|USFDC Home | USF Electronic Theses and Dissertations||| RSS|
This item is only available as the following downloads:
A Novel Biostable 3D Porous Collagen Scaffold for Implantable Biosensor by Young Min Ju A dissertation submitted in partial fulfillment of the requirement s for the degree of Doctor of Philosophy Department of Chemical & Biomedical Engineering College of Engineering University of South Florida Major Professor: Francis Moussy, Ph.D. Yvonne Moussy, Ph.D. Mark Jaroszeski, Ph.D. Michael VanAuker, Ph.D. Julie P. Harmon, Ph.D. Date of Approval: December 7, 2007 Keywords: implantable glucose sensor porous scaffold, NDGA crosslinking, microspheres, dexamethasone Copyright 2008, Young Min Ju
DEDICATION To my parents, who have always supported and encouraged me To my wife, Hee Jung, for always being there with devotion, patience and love To my son, Justin, a marvelous blessing
ACKNOWLEDGEMENTS Firstly, I would like to express my sincere gratitude to my dissertation advisor, Dr. Francis Moussy for his guidance, support, and encouragement in completing Ph.D. research project. Thank you for giving me the freedom to pursue this project and for developing my research career as a scientist. I will be forever grateful to you for your generosity, kindness, and wisdom over the years. I extend my appreciation to the members of my Ph.D. dissertation committee, Dr. Yvonne Moussy, Dr. Mark Jaroszeski, Dr. Michael VanAuker, Dr. Julie P. Harmon, who encouraged and leaded me to the right directions during my Ph.D. study. I would also like to thank Dr. Michael Weng, who served as an external committee chairperson for my dissertation defense. I would also like to thank the following for their assistance; Dr. Thomas J. Koob and Mr. Douglas Pringle for handling down th eir new technical knowledge; Ms. Margi Baldwin for her assistance with animal surgery; Ms. Sandy Livingston for her help with histology; Mr. Jay Bieber for his assistance with SEM analysis; Ms. Carla Webb, Ms. Cay Palez, Mr. Jamie Fargen, and Mr. Ed Van Etten for help with administrative and IT support. I thank all of my fellow in Biosensor & Biomaterials Lab.; Leigh West, Bobby Yu Bazhang, Nathan Long, Paul Dungel, Eric Guegan, Nuvala Fomban, Jose Rey, and James Merker, and in Dr. Harmons Lab.; Chunyan Wang, Moo Sung Kim, Kadine Mohomed. I would also like to thank Ms. Br ett Montegny for her tutor to improve my presentation skills. I would also like to thank Won-Seok, Chungsik, Man Soo, Seung Ryong, Byung Ryong, my Korean friends in the USF. I am also grateful to Dr. Hyung Bae Jung and Dr. In Ho Ra, visiting professor from Korea, whose generosity, advice and encouragement. Last but certainly not least, I especially thank my parents and parents-in-law who have always supported and encouraged me with unconditional charity and their prayer. I am grateful to my sister as well for her support. I am forever grateful to my wife Hee Jung. She is always by my side with devotion, patience, love, and constant cheers. Without her, I would not have completed this dissertation. Half percent of this Ph.D. dissertation belongs to her. I am also grateful to my sweet little son, Justin (A-rang), who is a blessing and a treasure of my heart.
NOTE TO READER The original of this document contains color that is necessary for understanding the data. The original dissertation is on file with the USF library in Tampa, Florida.
i TABLE OF CONTENTS LIST OF TABLES iv LIST OF FIGURES v LIST OF ABBREVIATIONS viii ABSTRACT x CHAPTER 1 INTRODUC TION 1 1.1. Diabetes 1 1.2. Implantable Glucose Sensor 3 1.3. Biocompatibilit y of Implanted Devices 7 1.4. Strategies for Biocom patible Implantable Sensors 14 1.5. Collagen and Its Us e in Biomaterials 17 CHAPTER 2 IN VITRO / IN VIVO STABILITY OF THE SCAFFOLDS AND IN VITRO SENSITIVITY OF IMPLANTABLE GLUCOSE SENSORS WITH SCAFFOLDS 21 2.1. Introduction 21 2.2. Materials and Methods 25 2.2.1. Materials 25 2.2.2. Preparation and Cross linking of Collagen Scaffolds 25 2.2.3. In vitro and In vivo Evaluation of Collagen Scaffolds 27 2.2.4. Preparation of Por ous Collagen Scaffolds around Implantable Glucose Sensors 28 2.2.5. In vitro Characterization of Sensors Coated with Scaffolds 30 2.3. Results and Discussion 32 2.3.1. Preparation of Porous Crosslinked Collagen Scaffolds 32 2.3.2. In vitro and In vivo Evaluation of Collagen Scaffolds 37 2.3.3. Porous Collagen Scaffolds around Implantable Glucose Sensors 40 2.4. Conclusions 48 CHAPTER 3 LONG-TERM IN VITRO / IN VIVO PERFORMANCE OF IMPLANTABLE GL UCOSE SENSORS WITH CROSSLINKED COLLAGEN SCAFFOLDS 49
ii 3.1. Introduction 49 3.2. Materials and Methods 52 3.2.1. Materials 52 3.2.2. Preparation of Po rous Collagen Scaffolds around Implantable Glucose Sensors 52 3.2.3. Long-term In vitro Characterization of Sensors Coated with Scaffolds 53 3.2.4. Implantation Procedures 55 3.2.5. Long-term In vivo Evaluation of Sensors Coated with Scaffolds 56 3.3. Results and Discussion 59 3.3.1. Preparation of Im plantable Glucose Sensors with Porous Crosslinked Collagen Scaffolds 59 3.3.2. Long-term In vitro Evaluation of Sensors with Porous Collagen Scaffolds 62 3.3.3. Long-term In vivo Performance of Sensors with Porous Collagen Scaffolds 64 3.4. Conclusions 73 CHAPTER 4 DEXAMETHASONE-LOADED PLGA MICROSPHERES/ COLLAGEN SCAFFO LD COMPOSITE SYSTEM FOR IMPLANTABLE GLUCOSE SENSORS 75 4.1. Introduction 75 4.2. Materials and Methods 78 4.2.1. Materials 78 4.2.2. Preparation of Dex-loaded Microspheres 78 4.2.3. Microsphere Analysis 79 4.2.4. Preparation of Dex-loaded Microspheres/ Scaffold Composite System 80 4.2.5. In vitro Release of Dex from Microspheres/ Scaffold Composite System 82 4.2.6. Preparatio n of Implantable Glucose Sensors with Microspheres/Scaffold Composite System 82 4.2.7. Implantation Procedures 83 4.2.8. In vivo Evaluation of Sensors Coated with Microspheres/Scaffold Composite System 84 4.3. Results and Discussion 86 4.3.1. Preparation of De x-loaded PLGA Microspheres 86 4.3.2. Preparation of Dex-loaded Microspheres/ Scaffold Composite System 90 4.3.3. In vitro Drug Release Studies 92 4.3.4. Implantable Glucose Sensors Covered with Microspheres/Scaffold Composite System 95 4.3.5. In vivo Performance of Sensors with Dex-loaded Microspheres/Scaffold Composite System 97
iii 4.3.6. Suppression of Inflammation to Dex-loaded Microspheres/Scaffold Composite System 103 4.4. Conclusions 106 CHAPTER 5 SUGGESTIONS FOR FUTURE STUDY 107 REFERENCES 108 APPENDICES 122 Appendix A: Protocol Prepar ation Procedure of Coil-type Glucose Sensors 123 A.1. Coiling of Platinu m-iridium (Pt-Ir) Wires 123 A.2. Enzyme Coating 123 A.3. Epoxy-PU Coating 124 Appendix B: Protocol Meas urement of Sensor Function 125 B.1. Preparation of Measurement 125 B.2. Response Time and Slope Measurement 125 B.3. Preparation of Calibration Plot 126 Appendix C: Protocol Implantation of Glucose Sensors in the Rat and Measurement of Sensor Function In vivo 128 C.1. Surgery Materials 128 C.2. Glucose Monitoring and Testing Apparatus 128 C.3. Sterilization and Pre-ca libration of Glucose Sensors 129 C.4. Protocol for Animal Surgery 129 C.5. Implantation of S ensors (Long Wire Sensors) 130 C.6. Sensors Testing 131 C.7. Animal Recovery 132 ABOUT THE AUTHOR End Page
iv LIST OF TABLES Table 1.1. Examples of App lications of Collagen-Based Medical Devices  18 Table 3.1. Number of Working Sensors after Implantation 68 Table 4.1. Solvent Effect on the Amount of Dex Load ing Efficiency and Encapsulation Efficiency 88 Table B.1. Changes of Glucose Concentration in the Cell 127
v LIST OF FIGURES Figure 1.1. Demonstration of Gluc ose Rise and Fall in Relation to Meals and Exercise 4 Figure 1.2. Schematic Illustration of the Needle-type Implantable Glucose Sensor 6 Figure 1.3. Temporal Variation in Tissue Reaction to Implanted Biomaterials 8 Figure 1.4. Schematic of Process of Wound Healing in the Presence of an Implant 11 Figure 1.5. SEM Photographs of Tips of Glucose Sensors 12 Figure 1.6. Light Micrograph Image of Glucose Sensor Tip after 10 Days of Implantation in Subcutis 13 Figure 2.1. Schematic Diagram of the Scaffold-coated Sensing Element of the Glucose Electrode 29 Figure 2.2. Schematic Mechani sm for (A) GA and (B) NDGA Crosslinking of the Collagen Scaffold 34 Figure 2.3. SEM Morphology of the Collagen Scaffold 35 Figure 2.4. Bulk Propert ies of GAand NDGA-crosslinked Scaffold 36 Figure 2.5. Collagenase Resistance of GAand NDGA-crosslinked Scaffold In vitro 38 Figure 2.6. SEM Morphology of the Scaffold after In vitro Degradation Study 39 Figure 2.7. In vivo Stability of GAand NDGA-crosslinked Scaffold in Rat Subcutaneous Tissue 41
vi Figure 2.8. Light Microscope Pictur es of the Implantable Glucose Sensing Element and SEM Morphology of the Scaffold Region 43 Figure 2.9. Amperometric Response Curves of the Glucose Sensors from 5 to 15 mM Glucose Concentration 44 Figure 2.10. Amperometric Response of Uncoated and Collagen Scaffold-coated Glucose Sensors (2-30 mM Glucose) 46 Figure 2.11. Effect of the Scaffold Th ickness on Glucose Sensor Sensitivity 47 Figure 3.1. Photograp h Showing (A) Long Wire and (B) Short Wire Collagen Scaffold-coated Glucose Sensors 54 Figure 3.2. Surgical Procedures by Two Different Implantation Techniques for Long Wire Sensors and Short Wire Sensors 57 Figure 3.3. Photographs of Implantable Sensors Coated with (A) GA-crosslinked Porous Collagen Scaffold and (B) NDGA-crosslinked Porous Collagen Scaffold 60 Figure 3.4. Schematic of Short Wire Implantable Glucose Sensor 61 Figure 3.5. Long-term In vitro Sensitivity Changes of Control Sensors and Sensors with NDGAor GA-crosslinked Collagen Scaffolds 63 Figure 3.6. Photograph of In vivo Continuous Glucose Monitoring Procedure 65 Figure 3.7. Long-term In vivo Sensitivity Changes of Control Sensors and Sensors with NDGAor GA-crosslinked Scaffold 67 Figure 3.8. Representative Photograph of Scaffolds In situ after 4 Weeks Post Implantation 70 Figure 3.9. Hematoxylin and Eosin Stained Sections Showing Tissue Surrounding Porous Scaffolds 71 Figure 4.1. SEM Morphol ogy of the Dex-loaded PLGA Microspheres 87
vii Figure 4.2. SEM Morphol ogy of the Dex-loaded PLGA Microspheres/Collagen Scaffold Composite 89 Figure 4.3. The Amount of Dex Loading in the Composite as Fabricated Using Either Water or Hydrogel Suspension with Different Initial Microspheres Loading Amounts 91 Figure 4.4. The Amount of De x Loading in the Composite as Fabricated Using Either Wa ter or Hydrogel Suspension after Rinsing with Water 93 Figure 4.5. Cumulative Dex Released from Standard Microspheres and Dex-loaded Microspheres/ Scaffold Composite During the In vitro Release Studies in PBS at 37C 94 Figure 4.6. Light Microscope Phot ographs of the Implantable Glucose Sensing Element with Dex-loaded Mi crospheres/Scaffold Composite 96 Figure 4.7. Effect of Adding PLGA Microspheres in the Scaffold on Glucose Sensor Sensitivity wit h Different Suspensions 98 Figure 4.8. In vivo Sensitivity Changes (Bar Graph results are shown as means SD) and Number of Working Sensors (Line Graph) of Control Sens ors and Sensors with NDGAor GA-crosslinked Collagen Scaffolds and Sensors with Dex-loaded Microspher es/NDGA-crosslinked Collagen Scaffold after 2 W eeks Post Implantation 99 Figure 4.9. Light Microscope Phot ographs of Implantable Glucose Sensors 101 Figure 4.10. Amperometric Response Curves of the Explanted Nonfunctioned Glucose Sensors after 4 Weeks Post Implantation 102 Figure 4.11. Hematoxylin and Eosi n Stained Sections of Tissue Surrounding Porous Scaffolds in Rats 104 Figure B.1. Amperometric Response Curve 126
viii LIST OF ABBREVIATIONS ADA American Diabetes Association SMBG Self-monitoring of blood glucose CGMS Continuous glucose monitoring system Ag/AgCl Silver/silver chloride GOD Glucose oxidase ECM Extracellular matrix FBC Foreign body capsule SQ Subcutaneous IV Intravascular IP Intraperitoneal PEG Poly(ethylene glycol) PHEMA Poly(hydroxyethylmethacrylate) PU Polyurethane TRM Tissue response modifiers PLGA Poly(lactic-co-glycolic acid) Dex Dexamethasone PVA Poly(vinyl alcohol) VEGF Vascular endothelial growth factor PDGF Platelet-derived growth factor GA Glutaraldehyde HMDI Hexamethylene diisocyanate EDC 1-ethyl-3-(3-dimethy laminopropyl) carbodiimide NHS N -hydroxysuccinimide NDGA Nordihydroguaiaretic acid PLLA Poly(L-lactic acid)
ix UV Ultra-violet BSA Bovine Serum Albumin Dc Degree of crosslinking S Sensitivity of glucose sensor ePTFE expended Poly(t etrafluorethylene) LCST Lower critical solution temperature HPLC High performance liquid chromatography
x A Novel Biostable 3D Porous Collagen Scaffold for Implantable Biosensor Young Min Ju ABSTRACT Diabetes is a chronic metabolic di sorder whereby the body loses its ability to maintain normal glucose levels Despite of developm ent of implantable glucose sensors in long periods, none of the biosensors are capable of continuously monitoring glucose levels during long-term implantation reliably. Progressive loss of sensor function occu rs due in part to biofouling and to the consequences of a foreign body response such as inflammation, fibrosis, and loss of vasculature. In order to improve the function and lifetime of impl antable glucose sensors, a new 3D por ous and bio-stable collagen scaffold has been developed to improve the biocompatib ility of implantable gluc ose sensors. The novel collagen scaffold was crosslinked using nor dihydroguaiaretic acid (NDGA) to enhance biostability. NDGA-treated collagen scaffolds were stable without any physical deformation in the subcutaneous tiss ue of rats for 4 weeks. The scaffold application does not impair the function of our sensor. The effect of the scaffolds on sensor function and biocompatibili ty was examined during long-term in vitro and in vivo experiments and compared with contro l bare sensors. The sensitivity of the short sensors was gr eater than the sens itivity of long sensors presumably
xi due to less micro-motions in the sub-cu tis of the rats. The NDGA-crosslinked scaffolds induced much less inflammation and retained their physical structure in contrast to the glutaraldehyde (GA)-crosslinked scaffolds. We also have developed a new dexamethasone (Dex, anti-inflammatory drug)-loaded poly(lactic-co-glycolic acid ) (PLGA) microspheres/porous collagen scaffold composite for implantable glucose sensors. The composite system showed a much slower and sustained drug release than the standard microspheres. The composite system was also shown to not significantly affect the function of the sensor s. The sensitivity of the sensors with the composite system in vivo remained higher than for sensor s without the composites (no scaffold, scaffold without microspheres). Histology showed that the inflammatory response to the Dex-loaded composite wa s much lower than for the control scaffold. The Dex-loaded composite syst em might be useful to reduce inflammation to glucose sensors and ther efore extend their function and lifetime.
1 CHAPTER 1 INTRODUCTION 1.1. Diabetes Diabetes is a chronic metabolic disorder in which the body loses its ability to maintain normal glucose levels. Diabetes is the 6th leading cause of death by disease and is rapidly increasing in the United States and around the world. The American Diabetes Association (ADA) estima tes that at least 20.8 million or 7% of Americans have suffered from diabetes, caused by a lack or shortage of insulin, the hormone that allows glucose to enter the bodys cells and be stored or used for physiological activation energy [1,2]. There are two major types of diabetes (Type I and II). Type I, or insulindependent diabetes, is an autoimmune dis ease. It is marked by blood sugar levels rising out of control becaus e the bodys immune system destroys the insulin-producing beta cells in the pancr eas. The pancreas then produces little or no insulin. Approximately, 510% of diabetes cases in t he US is Type I. Type II diabetes is the most common form of diabetes It is characterized clinically by hyperglycemia and insulin resistance, whic h results when the insulin produced, does not adequately control the uptake of glucose by the cells. Type II diabetes
2 is usually the type of diabet es diagnosed in patients that are over 30 years old or obese. Ninety percent of diabetes cases is Type II [2-4]. Diabetes has acute and chronic effects on the body, and may lead to death. Persistent abnormal high levels of blood glucose can slowly damage both the small and large blood vessels in the body, result ing in numerous complications [2,3], such as: Heart disease and stroke High blood pressure Blindness Kidney disease Nervous system disease Amputations Dental disease Complications of pregnancy Erectile dysfunction Thus, physicians and researchers are trying to develop better ways of monitoring and curing diabetes to avoi d life-threatening events.
3 1.2. Implantable Glucose Sensor The ADAs Consensus Statement on Self-Monitoring of Blood Glucose (SMBG) recommends that diabetic patient s should test their blood glucose level at least twice for Type II diabetes and four (for Type I diabetes) times a day . To maintain normal or near normal blood gl ucose levels (70-120 mg/dL), diabetic patients require injections of insulin, and have to monitor their own blood glucose levels throughout the day. Ho wever, the general use of over-the-counter glucose meters requires finger pricking to obtain blood samples several times each day. Because of the high density sensory neurons located in the dermis on the finger tip, patients frequent suffer fr om painful . Thus, the pai nfulness, inconvenience, and discomfort of self-monitoring of bl ood glucose are frequent obstacle to effective patient compliance and opt imal management of diabetes. To corrective regulate tight blood gluc ose control, a continuous glucose monitoring system (CGMS) is required. T he CGMS can provide additional data to track unpredictable glucose trend in relation to meals and exercise [Fig. 1.1] and allow hypoglycemic and hyperglycemic excu rsions to be avoided. During the past thirty years many kinds of continuous glucose monitoring systems have been studied. These include sensors implante d in the subcutaneous tissue [7-13], sensors implanted in th e vascular bed [14,15], and determining glucose concentration in interstitial fluid sampl ed using a micro dialysis device [16-18]. Although several studies of implantable glucose s ensors have been reported, none of the biosensors tested well capabl e of reliable in continuous blood
4 Figure 1.1. Demonstration of Gluc ose Rise and Fall in Relation to Meals and Exercise. Figure adapted from Joseph and Torjman .
5 glucose monitoring during long-term impl antation, progressive loss of sensor function occurred due in part to biofouli ng and to the consequences of a foreign body response, such as inflammation, fi brosis, and loss of vasculature [20-22]. Most of the implantable glucose sensors are based on amperometric enzyme sensors from the pioneering work of Clark and Lyons , Updike and Hicks , and Gough et al. . The typical enzyme-based amperometric sensor is composed of a two-electrode syst em with a glucose indicating platinum (Pt) working electrode and a silver/silver chloride (Ag/AgCl) reference-counter electrode. Figure 1.2 shows the needl e-type implantable glucose sensor commonly used for subcutaneous insertion . An outer layer of polyurethane membrane is permeable to glucose and oxygen but impermeable to most interfering substances. A crosslinked glucose oxidase (GOD) enzyme layer is sandwiched between inner and outer membrane. In the presence of oxygen, glucose is oxidized by GOD and produces hydrogen peroxide (H2O2). Hydrogen peroxide is then oxidized elec trochemically at the Pt electrode surface using a polarization voltage of about +700 mV, producing 2ethat is detected as a current [21,27]. The chemical reactions are: Glocose + O2 GOD Gluconic Acid + H2O2 H2O2 mV 700 2e+ 2H+ + O2
6 Figure 1.2. Schematic Illust ration of the Needle-type Im plantable Glucose Sensor. Figure adapted from Pickup et al. .
7 1.3. Biocompatibility of Implanted Devices Most implanted medical devices, including biosensors, frequently encounter a sequence of common host defense mechanisms, such as acute and chronic inflammation, wound healing, and fo reign body responses  [Fig. 1.3]. Acute inflammation begins within a few minutes after device implantation, with accumulation of interstitial fluid, plas ma proteins, and migration of leukocytes (neutrophils, monocytes, macrophages) ar ound sensors. Chronic inflammation follows if acute inflammation is not resolved. In general, macrophages rapidly differentiate from monocytes and become the predominant cell ty pe in exudates surrounding the devices. The macr ophages are key mediators in the development of immune reactions to impl anted synthetic biomaterials. They also produce and secrete a number of biologically active products including chemotactic factors, reactive oxygen met abolites, growth factors, and cytokines . Wound healing is the repair and remo deling process which occurs after. It takes place in the space between the im plant and the surrounding tissue. It is begun by the action of monocytes and macr ophages, followed by proliferation of fibroblasts and vascular endothelial cells at the wound site. The fibroblasts and new small blood vessels proliferate in developing granulati on tissue . The new small blood vessels are budded or sprouted from preexisting blood vessels. This process is called neovascularizati on or angiogenesis [30-32]. Fibroblasts also synthesize type III collagen and proteogl ycans at the wound site. Eventually, collagen deposition may result in the fo rmation of fibrous capsule around the implanted device.
8 Figure 1.3. Temporal Variation in Tiss ue Reaction to Implanted Biomaterials. Figure adapted from Anderson .
9 The foreign body reaction has a connection with foreign body giant cells and granulation tissue including macrophages, fibroblasts and new capillaries at the tissue-implant interface. Fibrosis or fibrous encapsulation is the end-stage of the healing process. Fibrotic tissue su rrounding the implanted device isolates it from the local tissue environment. Figur e 1.4 shows the process of wound healing in the presence of an implant. Pore size and pore density on the su rface of implanted device (i.e. scaffold) may greatly influence fibrous capsule thickness, blood vessel density, and the location of vessels within the th ree-dimensional scaffo ld . Large pore scaffolds (pores > 8 microns in diamet er) allow deep penetration of capillaries and supporting extra-cellular matrix (ECM). Sharkawy et al.  showed that after four weeks of subcutaneous impl antation in rat, well-organized collagen capsule typical of foreign-body res ponses around non-porous implants, while porous implants produced less fibrosis and more vascularized fibrous capsules. For implantable biosensors, adsorption of proteins and cells as well as the formation of a fibrous capsule tissue ca n severely hinder transport of small molecules, i.e. glucose. Glucose is not able to freely diffuse from capillary blood to the sensors transducer surface . Pi ckup et al.  reported an example of protein and cellular accumulation on the tips of the non-functioned glucose sensors after only five hours of implantat ion [Fig. 1.5]. Ertefai and Gough  showed fibrous capsule tissue surrounding a glucose sensor tip after 10 days of implantation in subcutis [Fig. 1.6].
10 Reichert and Sharkawy  reviewed the findings of several implantable biosensor studies: Inflammatory cells bind to and degrade sensor performance. Protein adsorption hinders sensor f unction by lowering permeability to glucose and oxygen. Fibrous tissue and exogenous pool of foreign body capsule (FBC) presents a transport barrier to glucose. Vascularization of the FBC is necessary for good long-term stability of response. Sensors inactivated in vivo often regain function when FBC is removed and retested in vitro Sensor baseline and sensitivity gr adually degrade with im plantation time. Sensor performance is erratic for t he first hour and then becomes steady upon equilibration. Subcutaneous (SQ) glucose levels lag behind plasma levels by 5-20 min. Intravascular (IV) implantation gi ves immediate glucose readings but suffers from thrombus formation. IV implantation is best if the sens or is placed in fast-moving blood stream. Intraperitoneal (IP) FBC is thinner than SQ. Textured coatings produce vascula rized FBC that might ensure longterm SQ sensor accuracy.
11 Figure 1.4. Schematic of Process of Wound Healing in the Presence of an Implant. Figure adapted from Cannas et al. .
12 B C A B C A Figure 1.5. SEM Photographs of Tips of Glucose Sensors. (A) Control sensornot implanted; (B) Functi oning sensor showing minimal biofouling; (C) Non-functioning se nsor showing significant protein and Cellular accumulation. Figure adapted from Pickup et al. .
13 Figure 1.6. Light Micrograph Image of Glucose Sensor Tip after 10 Days of Implantation in Subcutis. Note dense fibrous capsule surrounds sensor. Figure adapted from Pickup et al. .
14 1.4. Strategies for Biocompa tible Implantable Sensors Many researchers studied sensor modification to reduce sensor membrane biofouling in vivo One approach is to reduc e protein adsorption. Quinn et al.  used poly(ethylene gl ycol) (PEG) into a poly(hydroxyethylmethacrylate) (PHEMA) for surface modi fication of the biosensor. The PEG chains tend to stand perpendicular to t he membrane surface to provide a water rich phase that resists many protein mole cules. Vadgama et al [36,37] tried to reduce protein adsorption by using di amond-like carbon, so-called inert materials. Shichiri et al.  incorpor ated an alginate/polylysine gel layer on the sensor. Shaw et al.  reported bioc ompatibility improvem ent of biosensor, coated with PHEMA/polyurethane (PU). W ilkins et al.  and Moussy et al. [7,41-43] introduced the NafionTM (perfluorosulphonic acid) membrane, to reduce biofouling on surface of the sensor and reduce inte rference from urate and ascorbate. Armour et al.  coated their sensor ti ps with crosslinked albumin and Kerner et al.  dev eloped cellulose-coated sens ors to improve sensor blood compatibility. Controlled delivery of tissue response modifiers (TRM) can be used to control tissue responses. Dexamethasone (Dex), a synthetic glucocorticoid, is well known for its immunosuppressive and anti-inflammatory function [45-47]. The biosensor design could incorporate th is anti-inflammatory agent, which could be slowly released using biodegradable microspheres [48,49]. Typically, microspheres are prepared us ing natural or synthetic biodegradable polymers such as poly(lactic-co-glycolic acid) (PLGA) . Moussy et al [50,51] developed
15 Dex/PLGA microspheres designed to suppr ess the inflammatory tissue response to an implanted biosensor. Norton et al.  and Patil et al.  modified hydrogel coatings [PHEMA and poly(vinyl alcohol) (PVA) hydrogel, respectively] to include Dex-loaded PLGA microspher es to improve implantable biosensor biocompatibility. The best tissue environment for an implantable biosensor is vascularized tissue around sensor. Angiogenesis, which include as complex cascade of events involving endothelial cell activa tion, migration and proliferation, organization into immature vessels, associ ation of mural cells with the immature vessels, and matrix deposition as the vessels mature [54,55], has been extensively studied. The control of neova scularization has recently focused on the use of angiogenic growth factors such as vascular en dothelial growth factor (VEGF) and platelet-derived gr owth factor (PDGF). VEGF is a specific mitogen for initiating angiogenesis, specifically for promoting vascular permeability, prolifieration, and migration of endothel ial cells . PDGF promotes the maturation of blood vessels by the recr uitment of smooth muscle cells to the endothelium lining of nascent vasculature [55,57]. The controlled release of VEGF and PDGF has been studied widely as a strategy for increasing the blood vesse l density surrounding implants [58-61]. Klueh et al. [62,63] developed an in vivo gene transfer system with VEGF and found that the VEGF-biosens or systems induced neovascularization surrounding the sensor and thereby enhanced biosensor function in vivo Ward et al.  reported that VEGF infused continuously for 28 days into rat subcutaneous tissue
16 from a model biosensor led to local va scularization of the surrounding foreign body capsule. Norton et al.  modifi ed their hydrogel biosensor coatings to incorporate PLGA microspheres in order to release vascular endothelial growth factor.
17 1.5. Collagen and Its Use in Biomaterials In recent years, collagen and its der ived matrices have become the most widely used natural polymers in the biom edical field including tissue engineering due to their low antigenicity, biodegradab ility and good mechanical, hemostatic and cell-binding properties [65-69]. A br oad range of potentially manufactured products based on collagen is covering many medical disciplines  (Table 1). Collagen is a major protein of connecti ve tissues in animals as well as a key structural component of the extracelluar matrix. It is distributed in skin, bones, teeth, tendons, eyes and mo st other tissues and organs [71,72]. The collagen molecule is a rod-like structure with a molecular weight of about 300,000 which forms a unique triple-helix configuratio n of three polypeptide subunits. Each collagen molecule is organized in a regular and hierarchical pattern forming fibrils and fibril bundles that result in a tough tissue [73,74]. The collagen family has been reported to contain at least 19 dist inct types. Among them, type I collagen is the most abundant in higher order anima ls in the skin, tendon, bone, and most collagenous tissue, while ty pe II is found in cartilage, and type III is found, together with type I, in skin, and blood vessels. Thus, type I collagen is predominantly encountered in biomaterials application as bioprosthetic devices and scaffolds [71,73]. In order to devise strategies for using collagen in the development of advanced biomaterials for bi omedical engineering, it is necessary to confer mechanical strength and enzymatic degrada tion (e.g. collagenase) resistance by introduction of chemical or physical cro sslinking into the molecular structure.
18 Table 1.1. Examples of Applications of Collagen-Ba sed Medical Devices . Medical Area Application Cardiovascular surgery Dentistry Dermatology General surgery Neurosurgery Ophthalmology Orthopedics Otology Urology Wound management Vessel replacement, heart valves Periodontal attachment, al veolar ridge augmentation Tissue augmentation Hernia repair, adhesion barriers, tissue adhesives Nerve conduits, nerve repair Corneal graft, vitreous replacement Bone repair, cartilage and ligament reconstruction Tympanic membrane replacement Ureter replacement, renal r epair, urinary incontinence Dressings
19 There are several methods for cross linking collagen-based biomaterials. Glutaraldehyde (GA) is the most widely used as a crosslinking agent for collagen-based biomaterials [ 65,75]. At neutral pH, GA reacts with amino groups and with other functional group in protei n, including carboxy and amide group . However, GA induces cytotoxicity in vivo, caused by the presence of unreacted residual groups or the release of monomers of small polymers during enzymatic degradation [77,78 ]. To avoid cytotoxicity and calcification of GAcrosslinked collagen, polyepoxy compou nds, including glycol and glycerol polyglycidyl ethers, have been examined as potential collagen crosslinking agents [79,80]. Polyepoxy compounds react with the free amines of lysine side chains on neighboring proteins. The he xamethylene diisocyanate (HMDI), homobifunctional reagent, has the ability to crosslink collagen via its lysine side chains. Chvapil et al. [ 81,82] reported that HMDI is an effective method for crosslinking of collagen and does not leave residues after crosslinking process. Crosslinking with carbodiimi de, 1-ethyl-3-(3dimethylaminopropyl) carbodiimide (EDC) and N -hydroxysuccinimide (NHS) being the most widely used as crosslinking agents, has the main advan tage in that it only facilitates the formation of amide bonds between am ino group on the collagen molecules without becoming part of t he actual linkage . This method provides good biocompatibility and higher cellular different iation potential [66,83,84]. Koob et al. [85-88] has newly developed a process for polymerizing ty pe I collagen fibers with nordihydroguaiaretic acid (NDGA), a plant -derived compound. NDGA crosslinking is effective at significant ly improving the mechanical properties of
20 synthetic collagen fibers. Also, NDGAcro sslinked collagen fiber s did not elicit a foreign body response nor did t hey stimulate an immune reaction in vivo during a six week implantation peri od. In addition, various physical treatments including ultra-violet or gamma-ray irradiati on, and dehydrothermal treatment, have been effectively used for introducing crosslinks to collagen matrices [89-92].
21 CHAPTER 2 IN VITRO / IN VIVO STABILITY OF THE SCAFFOLDS AND IN VITRO SENSITIVITY OF IMPLANTABLE GL UCOSE SENSORS WITH SCAFFOLDS 2.1. Introduction To maintain near normal blood glucos e levels (70-120 mg/dL), diabetic patients widely use over-the-counter gl ucose meters, which require finger pricking to obtain blood samples several times a day. The pain , inconvenience, and discomfort of self-monitoring of bl ood glucose (SMBG) are frequently obstacles to effective patient compli ance and optimal m anagement of diabetes. During the past 20 years many kinds of continuous glucose monitoring systems have been studied including sensors implanted in the subcutaneous tissue [7-13], sensors implanted in the vascular bed [14,15], and determining glucose concentration in interstitial fluid sampl ed using a micro dialysis device [16-18]. Although several studies of implantable glucose sensors have been reported, none of these biosensors are capable of cont inuously monitoring glucose levels during long-term implantation reliably. Progr essive loss of sensor function occurs due in part to biofouling and to the cons equences of a foreign body response such as inflammation, fibrosis and loss of vasculature [20-22].
22 Many researchers have modified the surface of the sensors to reduce membrane biofouling in vivo In an approach to reduce protein adsorption, Quinn et al.  used poly(ethy lene glycol) (PEG) in a pol yhydroxyethylmethacrylate (PHEMA)matrix. Since the PEG chains tend to stand up perpendicular to the membrane surface, they provide a water-ri ch phase that resists binding of many protein molecules. Vadgamas et al. [36, 37] reduced protein adsorption by using diamond-like carbon, so-called inert materi als. Shichiri et al.  incorporated an alginate/polylysine gel layer at the sensor. Shaw et al.  reported improvement in biocompatibility of a biosensor coated with PHEMA/PU (polyurethane). Wilkins et al.  and Moussy et al. [7,41-43] introduced NafionTM (perfluorosulphonic acid) membrane, to redu ce biofouling on the surface of the sensor and reduce interference from urat e and ascorbate. Armour et al.  coated their sensor tips with crosslinked albumin and Kerner et al.  developed cellulose-coated sensors to improve sens or blood compatibil ity. However, none of these approaches has been successful for long term, stable glucose monitoring. Collagen and its derived matrices ar e used extensively as natural polymers in the biomedical field includi ng tissue engineering due to its low antigenicity, its biodegr adability and its good mechanical, haemostatic and cellbinding properties [65-69]. In order to dev ise strategies for using collagen in the development of advanced biomaterials for biomedical engineering, it is necessary to confer mechanical st rength and resistance to enzymatic (collagenase) degradation resistance with chemical or physical crosslinking
23 strategies. There are several strat egies for crosslinking collagen-based biomaterials. Glutaraldehyde (GA) is t he most widely used as a crosslinking agent for collagen-based biomaterials [ 65,75]. However, GA and its reaction products are associated with cytotoxicity in vivo due to the presence of crosslinking byproducts and the release of GA-linked collagen peptides during enzymatic degradation [77,78]. To avoid in vivo cytotoxicity and subsequent calcification of GAcrosslinked collagen, several alternat ive compounds have been examined as potential collagen crosslinking agents [79,80] such as polyepoxy, hexamethylene diisocyanate (HMDI), 1-ethyl-3-(3-dim ethylamino-propyl)ca rbodiimide (EDC), and ultra-violet (UV) or gamma-ray irradiation. Koob et al. [85-88] recently described a process for crosslinking of type I coll agen fibers with nordihydr oguaiaretic acid (NDGA), a plant com pound with antioxidant properties. They showed that NDGA significantly improved the mechanical proper ties of synthetic collagen fibers. In addition, they showed that NDGA-cross linked collagen fibers did not elicit a foreign body response nor did they st imulate an immune reaction during six weeks in vivo The extent of crosslinking and choice of crosslinking agent may also affect the porosity and pore size of the scaffo ld and may greatly influence fibrous capsule thickness, blood vessel density, a nd the location of vessels within the three-dimensional porous scaffold . Large pore scaffolds (greater than 60 micron pore size) allow deep penetration of capillaries and supporting extracellular matrix (ECM). Sharkawy et al.  recent ly showed that after four
24 weeks of subcutaneous implantation in rat, a well-organized collagen matrix typical of a foreign-body response enc apsulated non-porous im plants, while the porous polyvinyl alcohol (PVA) implants produced less fibrous and vascularized tissue capsules. The goal of this study was to develop a new porous collagen scaffold around implantable glucose sensors for im proving their biocompatibility. We fabricated porous collagen scaffolds by us ing a freeze-drying method followed by crosslinking using NDGA or GA. We evaluated the resistanc e of NDGAand GAcrosslinked collagen scaffolds to degradation using both in vitro and in vivo experiments. We also applied the scaffolds around a coil-type implantable glucose sensor and measured sensor function in vitro
25 2.2. Materials and Methods 2.2.1. Materials Type I collagen (purified from feta l bovine tendon) was a generous gift from Shriners Hospital for Children (T ampa, FL). Nordihydroguaiaretic acid (NDGA) was purchased from Cayman Chemical Co. (Ann Arbor, MI). Glucose, bovine serum albumin (BSA) and 50% (w/w ) glutaraldehyde (GA) were obtained from Fisher Scientific (Pittsburgh, PA). Glucose oxidase (GOD) (EC 22.214.171.124., type X-S, Aspergillus niger 157,500 U/g), epoxy adhesive (ATACS 5104), polyurethane (PU), tetrahydrofuran (THF ) and collagenase (E C 126.96.36.199, type I, from Clostridium histolyticum, 302 U/mg) were obtained from Sigma-Aldrich (St. Louis, MO). Sprague-Dawley out-bred ra ts (male, 375-399 g) were purchased from Harlan (Dublin, VA). 2.2.2. Preparation and Crosslinking of Collagen Scaffolds The collagen scaffolds were prepar ed by a freeze-drying method. Collagen was dissolved in 3% acetic acid to prepare a 1% (w/v) solution. The solution was applied to a cyli nder-shaped polypropylene mold ( 10 mm, height 8 mm) and then freeze-dried. A cylindrical 3D porous scaffold was obtained. The scaffolds were then crosslinked with NDGA or GA to minimize solubility and improve resistance to collagenase degradation. For NDGA crosslinking, dried collagen scaffolds were briefly soaked in absolute ethanol, followed by soaking in 2 M of NaCl solution for 12 h at room temperature. Scaffolds were re-s uspended in oxygen sparged phosphate
26 buffered saline (PBS, 0.1 M NaH2PO4, pH 9.0) for 30 min. at room temperature. Scaffolds were then treated with 3 mg of NDGA in 1 mL of PBS as follow: NDGA was dissolved in 0.4 N NaOH at a concentra tion of 30 mg/mL. On e milliliter of the NDGA solution was added directly to PBS in which the scaffolds were suspended to a final concentration of 3 mg/mL. The scaffolds were agitated in the NDGA solution for 24 h at room temperature. T he scaffolds were removed, briefly rinsed with water and freeze-dried. For a comparative study of the effe ctiveness of the NDGA treatment, other scaffolds were treated with 0.5% GA for 2 h or 12 h in ethanol solution at room temperature. To prevent the dissolution or loss of the ma trix during the GA crosslinking process, we used 100% ethanol instead of water. The crosslinked scaffolds were washed with de-ionized water and freeze-dried again. The morphology of the scaffolds before/afte r crosslinking was examined using scanning electron microscopy (SEM) after go ld sputter coating of the samples in a metal evaporator according to standard procedures. To evaluate the stability of the scaffo ld after crosslinking, the degree of crosslinking (Dc) was estimated by weig hing the dried samples before and after crosslinking. Dc was calculat ed using the follo wing equation: Dc [%] = (sample mass after crosslinking / sample mass before crosslinking) 100
27 The swelling property of the porous scaffolds was examined by measuring water absorption. The scaffolds we re weighed after thorough drying (Wdry) and immersed in purified water. After 24 h, the scaffolds were removed from the water and immediately weighed again (Wwet). Water absorption was calculated by using the follo wing equation: Water absorption (%) = [(Wwet Wdry)/Wwet] 100 2.2.3. In vitro and In vivo Evaluation of Collagen Scaffolds To examine the biological stability of the crosslinked scaffolds, we performed in vitro and in vivo biodegradation tests. In vitro biodegradation of NDGAand GA-crosslinked scaffolds was tested using bacterial collagenase. Fabricated NDGAand GA-crosslinked co llagen scaffolds were incubated in the collagenase solution (1 mg/mL in PBS at 37C) for up to 4 weeks. Scaffolds were removed from the solution, rinsed with deionized water and freeze-dried at given time intervals (weeks 1 to 4) during incubation. The in vitro degradation was evaluated as the percentage of weight difference of th e dried scaffold before and after enzyme digestion. To determine the stability of the crosslinked scaffolds in vivo we directly implanted NDGAand GA-crosslinked collagen scaffolds in rats. The scaffolds were disinfected with 70% ethanol solu tion for 2 h and implanted subcutaneously in the back of the rats. Scaffolds were explanted at 7, 14, 21, and 28 days after implantation. After explant ation, the scaffolds were examined macroscopically.
28 2.2.4. Preparation of Porous Collagen Scaffolds around Implantable Glucose Sensors We first fabricated coil type glucose sensors loaded with crosslinked enzyme (GOD: Glucose Oxidase) using a Platinum-Iridium (Pt/Ir) wire (Teflon coated, 0.125 mm, Pt:Ir = 9:1, Medwire, Sigmund Cohn Corp.). Then, we applied bovine tendon type I collagen scaffolds around the sensors [Fig. 2.1]. Briefly, in order to fabric ate a glucose sensor, the Te flon coating of the top 10 mm of a Pt/Ir wire was removed and the wire was wound up along a 30-gauge needle to form a coil-like cylinder. The cylinder unit had an outer diameter of 0.55 mm and an inner diameter of 0.3 mm and a length of 1 mm. A cotton thread was inserted inside the coil chamber to re tain the enzyme solution during enzyme coating of the electrodes. GOD was added and crosslinked to the sensors by dip coating in an aqueous solution containi ng 1% GOD, 4% BSA, and 0.6% (w/w) glutaraldehyde. The outer membrane of the sensor was coated with EpoxyPolyurethane (Epoxy-PU) by dipping in Epoxy-PU solu tion (2.5% (w/v) in THF, Epoxy:PU = 1:1). The sensor was dried at room temperature for at least 24 h. The two ends of the sensing element were sealed by electrically-insulating sealant (Brush-On electrical tape, North American Oil Company) [93,94]. To apply collagen scaffolds around the sensors, the sensors were dipcoated with 1% (w/v) collagen solution and freeze-dried. The porous scaffolds around the glucose sensors were cross linked with either NDGA or GA as previously described. Obtai ned sensors were st ored dry at room temperature or
29 Figure 2.1. Schematic Diagr am of the Scaffold-coated Sensing Element of the Glucose Electrode. (1) Teflon-covered Pt-Ir wire; (2) Ag/AgCl reference wire; (3) Collagen scaffold; (4) Electrically-insulating sealant; (5) Epoxy-Pu outer membrane; (6) Enzyme layer; (7) Stripped and coiled Pt-Ir wire; (8) Cotton fiber with GOD gel.
30 in PBS at 4C. The morphology of t he sensors was observed using light microscope and SEM. In addition, in order to evaluate sensitivity changes of the sensors with varying wall thickness of the scaffold around the sensors, we controlled the wall thickness of the scaffold by multiple dipping /freezing cycles in collagen solution. The scaffold with t he sensor was then freeze-dried and crosslinked as previously described. Silver wires (Teflon coated, 0.125 mm, World Precision Instruments, Inc.) were used to fabricat e the Ag/AgCl reference elec trodes. Silver wires were coiled and anodized galvanostatically at 1mA overnight in stirred 0.1 M HCl [93,94]. 2.2.5. In vitro Characterization of Sensor s Coated with Scaffolds The glucose sensors were characterized in PBS (pH 7.4) at 700 mV versus the incorporated Ag/AgCl refer ence electrodes. The working electrode (Pt/Ir wire) and Ag/AgCl reference electr ode of each sensor were connected to an Apollo 4000 potentiostat (W orld Precision Instrument s, Inc.). The background current was allowed to stabilize for 10 min., and the sensors were then exposed to a series of glucose solutions in order to examine their sensitivities and linearities. The response sensitivity (S) was repeatedly assessed by 1) measuring the response current (I1) of a C1 glucose solution, 2) adding a concentrated glucose solution into the m easured solution to increase the glucose concentration to C2, and 3) measuring the response current (I2) of the resulting
31 solution. The sensitivity was expressed as the current increase caused by a 1 mM glucose increase, i.e. S = (I2 I1) / (C2-C1).
32 2.3. Results and Discussion 2.3.1. Preparation of Porous Crosslinked Collagen Scaffolds The chemistry of the NDGA crosslinking reaction differs from the reaction using the GA treatment [F ig. 2.2]. GA is the most common crosslinking agent used for fixation of collagen scaffolds fo r tissue bioengineering. Both aldehyde functional groups of the GA molecule react with amine groups between two neighboring polypeptide chains particularly lysine side chains. Unfortunately, GA crosslinking is encumbered with potentia l cytotoxicity problems caused by the presence of unreacted residual groups or the release of monomers and small polymers during enzymatic degradation [77,78]. NDGA treatment is an al ternative crosslinking agent, which possesses reactive catechols. Colla gen crosslinking with NDGA mimics the quinine tanning mechanism in the skate egg capsule. Catechol-quinone t anning systems are prevalent in a wide variety of animals, which the process serves to strengthen vulnerable extracellular matrices (e.g. insect cuticle, mussel byssus threads) [85,95]. NDGA, isolated from the creosote bush, is a low molecular weight dicatechol containing two ortho -catechols. The two catechols on NDGA undergo auto-oxidation at neutral or alkaline pH producing reactive quinones. Two quinones then couple via ar yloxy free radical formation and oxidative coupling, forming bisquinone crosslinks at each end. The NDGA continues forming a large crosslinked bisquinone polymer network in which the collagen fibrils are embedded. The NDGA treatment does not form crosslinks with amino acid side chains of collagen [85,86,95].
33 In this study, highly porous collagen scaffolds were prepared by a freezedrying method. We ascertained that t he obtained scaffolds have an open cell and interconnected pore structure based on SEM observation [Fig. 2.3(A)]. The pores of the scaffolds are regularly di stributed and range from 20 to 100 m in diameter (mean ~ 60 m). Sharkawy et al.  reported that the a 60 m mean-pore-sized polyvinyl alcohol (PVA) sponge provi ded a tissue in-growth environment and allows to infiltration of neovasculature but did not allow for fibrous tissue ingrowth. After crosslinking with NDGA and GA the pore size and pore structure of both scaffolds were not significantly altered [Fig. 2.3(B) and (C)]. Figure 2.4 shows the degree of crosslinking and wa ter absorption of the scaffolds using different crosslinking methods. The mass was reduced to about 70% after NDGA treatment and 60% for GA treatment after the crosslinking process due to the loss of uncrosslinke d collagen components. Crosslinked collagen scaffolds had significantly high er form stability than uncrosslinked collagen scaffolds. Also, the swelling behavior of NDGAand GA-crosslinked scaffolds showed no significant differences between the two different crosslinking agents. The water absorptions of both crosslinked scaffolds were above 99%. The high swelling property of sponge-lik e matrices seems to be dependent on the porous inner structur e of the scaffold, which possesses good absorbent characteristics .
34 NH2 C H C H O O N HC HC N NH2 + Collagen Glutharaldehyde Collagen + A HO HO OH OH CH3CH3 O O O O CH3CH3 Oxidation O O O O O O O O O O Free Radical Generation Oxidative CouplingBNDGA (di-catechol)Di-quinone Quinone Bisquinone GA cross-linked Collagen O O O O O O O O O O O O O O O O O O O O NDGA cross-linked CollagenCollagenCollagenCollagen Collagen Figure 2.2. Schematic Mechanism for (A) GA and (B) NDGA Crosslinking of the Collagen Scaffold.
35 A B C A B C Figure 2.3. SEM Morphology of the Co llagen Scaffold. Determination of the pore size of collagen scaffolds by SEM. (A) No crosslinking; (B) GA-crosslinked; (C) NDGA-crosslinked.
36 50 60 70 80 90 100 GA cross-linked Degree of Cross-linking Water AbsorptionNDGA cross-linkedDegree of Cross-linking (%)99.0 99.1 99.2 99.3 99.4 99.5 Water Absorption (%) Figure 2.4. Bulk Properties of GAand NDGA-crosslinked Scaffold. Results are shown as means SD (n=3).
37 2.3.2. In vitro and In vivo Evaluation of Collagen Scaffolds The biological stability of the crosslinked collagen scaffolds was investigated by in vitro and in vivo biodegradation tests. Degradation in both uncrosslinked (control) and crosslinked scaffolds was characterized by determining weight loss of the scaff old after enzymatic digestion. The uncrosslinked scaffolds and scaffolds crosslinked with GA for 2 hours were completely degraded in the collagenase solution within several hours while NDGAor GA-crosslinked (for 12 h) sca ffolds were not degraded within 24 hours. A significant increase in resistance to enzymatic digestion could be shown after crosslinking. Figure 2.5 s hows long-term collagenase in vitro degradation test (weight remaining %) of the NDGAand GA-crosslinked scaffolds. After 1 week exposure to collagenase, both types of scaffolds showed high resistance to enzymatic digestion (> 80% weight rema ining). After 3 and 4 weeks, all scaffolds retained 70% of their initial mass. Howe ver, in the case of GA-crosslinked scaffold, the pore size was increased after 4 weeks collagenase digestion process [Fig. 2.6(B) vs Fig 2.6(D)]. In contras t, the pore size of NDGAcrosslinked scaffolds did not appear to incr ease [Fig. 2.6(A,C)]. As a result, we suggest that NDGA or GA treatment can provide collagen scaffolds with improved enzymatic biodegradation stabilit y. The collagenase cleavage sites were more effectively blocked by the cr osslinking of the collagen scaffolds . To study the stability of the crosslinked scaffolds in vivo we implanted crosslinked collagen scaffolds in the subcutaneous tissue of the Sprague-Dawley
38 1234 0 20 40 60 80 100 % of Original WeightCollagenase Treatment Time (weeks) GA cross-linked* NDGA cross-linked Figure 2.5. Collagenase Resistance of GAand NDGA-crosslinked Scaffold In vitro Results are shown as means SD (n=3). Scaffolds were treated with 0.5% GA for 12 h in ethanol solution at room temperature.
39 C D A B Figure 2.6. SEM Morphology of the Scaffold after In vitro Degradation Study. Results are shown as means SD (n=3). (A) NDGA-crosslinked scaffold after 2 weeks collagenase treatment; (B) GA-crosslinked scaffold after 2 weeks collagenase treatment; (C) NDGAcrosslinked scaffold after 4 wee ks collagenase treatment; (D) GAcrosslinked scaffold after 4 weeks collagenase treatment.
40 rats and explanted samples two and four w eeks post implantation. After 2 weeks implantation, the NDGA-crosslinked scaffolds did not show evidence of degradation, but the ov erall shape of the GA-crosslinked scaffolds was deformed and the size slightly reduced because of st arting degradation [F ig. 2.7 (A)]. After 4 weeks, the size and shape of the GA-c rosslinked scaffolds were dramatically changed (-78% in size of 2 weeks)but t here was a small change in the NDGAcrosslinked scaffolds (-18.9% in size of 2 we eks) [Fig. 2.7(B)]. This indicated that the scaffolds treated with t he NDGA were more stabl e than the scaffolds crosslinked with the GA treatment used in these studies. 2.3.3. Porous Collagen Scaffold s around Implantable Glucose Sensors We first fabricated coil type glucose sensors loaded with crosslinked enzyme (GOD: Glucose Oxidase) by using Platinum-Iridium (Pt/Ir) wires. Then, we applied bovine tendon type I collagen sc affolds around the sensors [Fig. 2.1]. Yu et al.  previously reported that this coil-type sensor allows more GOD loading, provides a larger electrochemical surface ar ea, and therefore increases the response current as compared to a needle-type sensor. Our sensor is flexible and miniaturized (0.5 mm dia. ) for subcutaneous implantation. It is composed of a two-electrode system with a glucose indicating platinum electrode and a Ag/AgCl reference-counter electrode. Our sensor utilizes a three-layer membrane configuration of crosslinke d collagen scaffold, epoxy-polyurethane (Epoxy-PU) and GOD. The collagen scaffold (the outer layer in this case) can
41 NDGA cross-linked NDGA cross-linked A GA cross-linked GA cross-linked NDGA cross-linked B GA cross-linked NDGA cross-linked GA cross-linked Figure 2.7. In vivo Stability of GAand NDGA-crosslinked Scaffold in Rat Subcutaneous Tissue. (A) 2 weeks after implantation; (B) 4 weeks after implantation.
42 uptake 99% of its dry weight of water in cluding glucose and other molecules. The Epoxy-PU membrane under the scaffold is permeable to glucose and oxygen but impermeable to most interfering subs tances. GOD immobilized in a BSA/GA matrix is sandwiched between the Pt/Ir wire and the Epoxy-PU membrane. In order to eliminate air bubbles entrapped in the cham ber during coating, to stabilize the enzyme gel in side the chamber, and to ma ke the enzyme solution easier to remain in the coil, we used a cotton fiber inside the coil chamber. The collagen scaffolds were prepared by a freez e-drying method and crosslinked to minimize water solubility and enzymatic collagenase degradation. With a light microscope, we confirm ed that the porous scaffolds thoroughly surrounded the sensor tips [Fig. 2.8(A) and (B)]. We also observed the surface and crosssectional morphology of the scaffold s around the sensors using SEM. Many collagen fibrils and uniform open pore stru cture were observed on the surface [Fig. 2.8(C)]. Inter-connected open pores in the scaffold and a thickness of 150 200 m were observed in cross-sectional region [Fig. 2.8(D)]. The amperometric response curves of the glucose sensors with and without scaffold (control) were obtained by varying the glucose concentration from 5 mM to 15 mM as shown in Figure 2. 9. These glucose concentrations were selected because these concentrations were located in the linear response region (2 30 mM) of the studied sensors. The result s showed no significant response current change before and after sc affold application around the sensor. However, the sensors with scaffolds had a slower response time to reach equilibrium current (T95%) than control sensors. The response time, T95%, is
43 A B C D Figure 2.8. Light Micro scope Pictures of the Im plantable Glucose Sensing Element and SEM Morphology of the Scaffold Region. (A) Uncoated sensor; (B) Coated with sca ffold; (C) Surface; (D) Crosssection.
44 051015202530 100 150 200 250 300 Time (min)Current (nA) 3 2 1 T 95% =14.0 min T 95% =17.0 min T 95% =17.9 min Figure 2.9. Amperometric Response Curves of the Glucose Sensors from 5 to 15 mM Glucose Concentration. (1) Uncoated sensor; (2) Coated with GA-crosslinked scaffold; (3) Coated with NDGA-crosslinked scaffold. T95% is defined as the time at 95% of the maximum current change (I15 mM I5 mM).
45 defined as the time at 95% of the maximum current change (I2 I1). The T95% of control sensor was 14.0 min. whereas T95% of the sensors with NDGAand GAcrosslinked scaffold were 17.9 min. and 17. 0 min., respectively. The delay of the response time (17.9 and 17 min.) was probably caused by the added physical barrier of the porous scaffolds. The currents produced by sensors wit h NDGA-, GA-crosslinked scaffolds and without scaffolds in response to varyi ng glucose concentration (2 30 mM) are showed in Figure 2.10. The response currents of the control sensors in the high glucose concentration region (20 30 mM) were only a little higher than those of the sensors with sca ffolds. However, there was no statistical difference between control and sensors with scaffolds (p > 0.05; student t-test). The average sensitivity of the control, NDGAand GA-cro sslinked scaffold around sensors was 11.0, 7.1, and 8.1 nA/mM, respectively. Therefore, scaffold application around the glucose sensors did not negatively affect the function of the sensors. We also examined the sensitivity changes of the sensors with varying wall thickness of the scaffold controlled by di pping cycles in collagen solution. As can be seen in Figure 2.11, the s ensitivity of the 4 times dip-coated sensors remained at 60% of their initial sens itivity (no scaffold). When t he sensors were dip-coated more then 5 times, glucose could not di ffuse properly through the scaffolds. The sensitivity was dramatically reduced to below 20% of the in itial sensitivity. Although the porous scaffold material has good water absorbent properties, the wall thickness can affect the sensor function.
46 0102030 0 100 200 300 400 * * * ** ** ** ** ** ** ** Glucose Concentration (mM)Current (nA) Control (No scaffold) GA cross-linked NDGA cross-linked* Figure 2.10. Amperometric Response of Uncoated and Collagen Scaffoldcoated Glucose Sensors (2-30 mM Glucose). Results are shown as means SD (n=3). Indicates no st atistically significant differences between control and GA-crosslinked scaffolds at each glucose concentration (p > 0.05). ** Indica tes no statistically significant differences between control and NDGA-crosslinked scaffolds at each glucose concentration (p > 0.05).
47 034567 0 20 40 60 80 100 Control Change in Sensitivity (%)Number of Dipping CyclesIncreasing Scaffold Thickness Figure 2.11. Effect of the Scaffold Th ickness on Glucose Sensor Sensitivity. Results are shown as means SD (n=3).
48 2.4. Conclusions In this study, porous type I colla gen scaffolds were prepared by a freezedrying method then crosslinked using NDGA or GA treatments. The fabricated collagen scaffolds have an open cell and interc onnected pore structure. To allow the infiltration neovasculature but to rest rict fibrous tissue formation, the mean pore size was controlled to 60 m by controlling the concentration of the collagen solution. Both crosslinking methods did not significantly affect the scaffolds geometry and bulk properties. They also had a similar resistance property to the collagenase enzyme in vitro However, NDGA-crosslinked scaffolds were shown to be more stable in vivo In addition, we also applie d the highly porous NDGAcrosslinked scaffolds to our implantable gl ucose sensor as a potential approach for reducing biofouling and improving biocompatibility. The porous scaffold application did not significantly affect the function of t he glucose sensor. Therefore, the application of an NDGA-crosslinked co llagen scaffold might be a good candidate for improving the biocompatib ility of implantable biosensors. We plan to use this scaffold to enhance t he function and lifetime of implantable biosensors by providing a controlled lo cal environment around the sensors with the additional help of various drugs and growth factors.
49 CHAPTER 3 LONG-TERM IN VITRO / IN VIVO PERFORMANCE OF IMPLANTABLE GLUCOSE SENSORS WITH CROSSL INKED COLLAGEN SCAFFOLDS 3.1. Introduction Although many strategies for cont inuous glucose monitoring have been developed over the past 30 years, achi eving reliable and continuous glucose monitoring in vivo is still a very difficult task. Very often, implantable glucose sensors lose function after a re latively short period of time in vivo or become unreliable, despite having excellent in vitro performances including good selectivity, a high sensitiv ity, and a fast response time [33,62,98-100]. This loss of function is in part a c onsequence of protein adsorption, inflammation, fibrosis encapsulation, and loss of vasculature resulting from t he biofouling and the tissue trauma caused by the host respons e to the sensor and the surgical implantation [20,21,48]. Ultim ately, biofouling of the biosensor membrane very much influences glucose diffusion, leading to in vivo sensor failures [101,102]. Overall, few successful long term implantations of glucose sensors have been reported. Armour et al.  implanted 6 sensors in travascularly in dogs for up to 108 days. Three sens ors still functioned with no adherent clots and with the same in vitro calibration curves before and after ex plantation. Updike et al. 
50 telemetrically monitored glucose us ing 3 implanted sensors. The sensor response dropped during the initial period in vivo but then rose and stabilized until 42-94 days. The same gr oup (Gilligan et al. ) observed a stable foreign body capsule (FBC) around the Dacron or ePTFE velour shells of their implanted sensors. The sensors eventually failed because of enzymatic degradation or biofouling of the sensor membranes. Picku p et al.  showed that only 50% of sensors implanted in non-diab etic subjects responded in vivo Explanted sensors examined by scanning electron microscopy were coated by cells and proteins at the sensor tip. Shichiri et al.  and Ertefai et al.  reported that the in vivo lag time was increased, compared to the in vitro lag time. The increase was attributed to protein deposition and FB C tissue at the sensor tip. In order to minimize biofouling and to improve sensor function, many researchers have designed new sensors with modifications to the surface of the sensor outer membrane. Moussy et al. [41,42] introduced a new sensor with a needle-type geometry and a Nafion outermost layer. Quinn et al.  used a photo-crosslinkable copolymer containing 2-hydroxyethyl methacrylate (HEMA) and poly(ethylene glycol) ( PEG) as a sensor coating material. The results showed that the copolymer-treated electrodes induced much less fibrous tissue than control electrodes due to good biologi cal performance of the PEG material. In order to reproduce lipid characterist ics to mimic the cell surface membranes, and induce anti-thrombogenicity, Nishi da et al.  synthesized a phosphorylcholine (PC)-containing polymer which was applied as a sensor membrane and showed excelle nt biocompatibility.
51 Because of good swelling and viscoel astic properties and outstanding biocompatibility, many researchers us e hydrogels such as PEG hydrogel (Quinn et al. ), phenylboronic acid-based hydrogel (Lei et al. ), and polyacrylamide hydrogel (Fernandez et al ) as the outermost coating of glucose sensors. Recently, numerous stra tegies to control delivery of tissue response modifiers (TRM) have been reported. For example, Gifford et al.  used nitric oxide (NO) to downregulate m ediators of the inflammatory response and Norton et al.  characteriz ed VEGF and dexamethasone (Dex) delivery from sensor coatings. We recently reported the developm ent of new porous collagen scaffolds which were applied around implantable gl ucose sensors to improve their biocompatibility. We fabr icated porous collagen scaffolds by using a freezedrying method followed by crosslin king using NDGA or GA . In a continuation of this study, we ev aluated the sensitivity of sensors with ether NDGAor GA-crosslinked collagen scaffolds during long-term in vitro and in vivo experiments. We also fabricated two different l engths of sensors (long and short wires) in order to minimize scaffo ld damage and compared their function in vivo to evaluate the effects of micro-motion on the sensors.
52 3.2. Materials and Methods 3.2.1. Materials Type I collagen (purified from feta l bovine tendon) was a generous gift from Dr. Thomas Koob, Shriners Ho spital for Children (Tampa, FL). Nordihydroguaiaretic acid (NDGA) was purchased from Cayman Chemical Co. (Ann Arbor, MI). Glucose, bovine serum albumin (BSA) and 50% (w/w) glutaraldehyde (GA) were obtained from Fisher Scientific (Pittsburgh, PA). Glucose oxidase (GOD) (EC 188.8.131.52., type X-S, Aspergillus niger 157,500 U/g), epoxy adhesive (ATACS 5104), polyuret hane (PU), and tetrahydrofuran (THF) were obtained from Si gma-Aldrich (St. Louis, MO). Dextrose injection solution (50%, w/v) was obtained from Abbott Labor atories (North Chicago, IL). The FreeStyleTM portable glucometer was from TheraSense (Alameda, CA). SpragueDawley out-bred rats (mal e, 375-399 g) were purchased from Harlan (Dublin, VA). 3.2.2. Preparation of Porous Collagen Scaffolds around Implantable Glucose Sensors We fabricated miniature coil type glucose sensors loaded with crosslinked enzyme (GOD: glucose oxidase) using a platinum-iridium (Pt/Ir) wire (Teflon coated, 0.125 mm, Pt:Ir = 9:1, Med wire, Sigmund Cohn Corp., Mount Vernon, NY). We applied bovine t endon type I collagen scaffolds around the sensors. Scaffolds were then cross linked with NDGA or GA treatment as previously described  to minimize solubility and to improve resistance to enzymatic degradation in vivo Control sensors (without scaffolds) and sensors
53 with NDGAor GA-crosslinked scaffolds were equilibrated in phosphate-buffered saline (PBS, 0.1 M NaH2PO4, pH 9.0) for 2 days at r oom temperature prior to being used in vitro or in vivo The initial sensitivity of all sensors was measured in 5 and 15 mM glucose in PBS. Amperometric meas urements were performed at room temperature at 0.7 V vs Ag/AgCl. The working electrode (Pt/Ir wire) and the Ag/AgCl reference electrode of each sensor were connected to an Apollo 4000 potentiostat (World Precision In struments, Inc., Sarasota, FL). In order to investigate the effect of wire length on the s ensor function, we fabricated sensors with two different lengt hs; 10 and 30 mm [Fig 3.1]. Only the wires were of different length, the sensing elements remained identical. 3.2.3. Long-term In vitro Characterization of Sensors Coated with Scaffolds In order to exam ine the long-term in vitro sensitivity of sensors, uncoated (control) sensors, sens ors with NDGA-crosslinked collagen scaffolds and sensors with GA-crosslinked collagen scaffo lds (n=8 / group) were incubated in PBS at 37C for 4 weeks. At 7, 14, 21, and 28 days, each sensor was removed and tested in glucose solution. The sensitivity of the glucose sens ors was characterized in glucose/PBS (pH 7.4) at 700 mV versus the incor porated Ag/AgCl reference electrodes. The background current was allo wed to stabilize for 10 mi n., and the sensors were then exposed to a series of glucose so lutions in order to examine their sensitivities and linearities.
54 A B A B Figure 3.1. Photograph Showing (A) Long Wire and (B) Short Wire Collagen Scaffold-coated Glucose Sensors.
55 The response sensitivity (S) was repeat edly assessed by 1) measuring the response current (I1) of a C1 glucose solution, 2) adding a concentrated glucose solution into the measured solution to increase the glucose concentration to C2 and 3) measuring the response current (I2) of the resulting solution. The sensitivity was expressed as the curr ent increase caused by a 1 mM glucose increase, i.e. S = (I2 I1) / (C2-C1). 3.2.4. Implantation Procedures All implantable glucose sensors were disinfected using 70% ethanol and then placed in sterile PBS prio r to implantation. During the surgical procedure, a continuous flow gas anesthesia system wa s used to deliver 1.5 % isoflurane to the rats in 2.0 L/min. ox ygen flow. All protocols were approved by the University of South Florida Institutional Animal Care and Use Committee (IACUC). Fortyeight sensors (eight control short sensor s; CS, eight control long sensors; CL, eight NDGA-crosslinked scaffold aroun d short sensors; NS, eight NDGAcrosslinked scaffold around long sensor s; NL, eight GA-crosslinked scaffold around short sensors; GS, and eight GA-crosslinked scaffold around long sensors; GL) were implanted subcutaneousl y on the back of the rats. Each rat received two of one type of sensors. For long sensors, two 1.5 cm long l ongitudinal incisions were made 1.5 cm laterally to the dorsal midline, and 3-4 cm caudally from the neck. A subcutaneous pocket was created using blunt surgical scissors. A 14 ga. I.V. catheter was inserted subcutaneously toward the incision from the 4-5 cm lower
56 back region. The needle was withdrawn leav ing the cannula in the subcutaneous tissue. The sensor wires were carefully fed into the cannula through the incision [Fig. 3.2(A)]. The sensor was secured to the skin by passing a 3-0 Prolene suture through the small gap of the wound clip co vering the sensor wires and incisions closed using 3-0 Prolene. T he cannula was then withdrawn, leaving the sensor in the subcutaneous tissue. For short sensors, same-sized incisions were made and a subcutaneous pocket was also created using blunt su rgical scissors before implantation. However, the sensors were directly im planted through the incision without using a cannula [Fig. 3.2(B)]. The sensors we re secured to the skin and incisions closed using the same approach utilized for the long sensors. In addition, in order to evaluate the inflammatory response of the tissue around and cellular intrusion into the co llagen scaffolds, we directly implanted NDGAand GA-crosslinked sca ffolds (without sensors) in the rats. At set time intervals, tissue samples containing the scaffolds were excised and embedded in paraffin. Sections (5 m in thickness) were cut and stained with Mayers hematoxylin and eosin (H&E) stain. Stained sections were analyzed and photographed using an Olympus BX41 mi croscope (Olympus; Tokyo, Japan). 3.2.5. Long-term In vivo Evaluation of Sensors Coated with Scaffolds The sensitivity of each sensor wa s measured every seven days for up to 28 days or until there was no amperometri c response from the implanted sensor. During each measurement period, four rats were anesthetized using isoflurane
57 Figure 3.2. Surgical Procedures by Two Different Implantation Techniques for Long Wire Sensors and Short Wire Sensors. (A) Long wire sensor (using a 14 ga. catheter guidance) ; (B) Short wire sensors (direct implantation).
58 and the eight implanted sensors were cont inuously monitored using two Apollo 4000 potentiostats. After a stable signal wa s obtained from the sensors, 0.7 mL of sterile 50% dextrose was administer ed intraperitoneally using a 27 ga. needle. Following the injection, sma ll blood samples were collected every 7 minutes from the rat tail and the glucose level wa s determined using the standard FreestyleTM glucometer. The amperometric response co rresponding to the glycemia of the rat was recorded at the corresponding current -time intervals of each sensor. The sensor sensitivity was calculated by di viding the change in current (I) by the change in glycemia (C) between the initia l (before dextrose injection) and the peak status (after dextrose injection) as follows: Sensitivity (nA/mM) = (Imax I0) / (Cmax C0)
59 3.3. Results and Discussion 3.3.1. Preparation of Im plantable Glucose Sensors with Porous Crosslinked Collagen Scaffolds In order to create a porous scaffold for implantable glucose sensors, we first fabricated coil type glucose sensors loaded wi th crosslinked enzyme (GOD: glucose oxidase) using platinum-iridi um (Pt/Ir) wires. Then, we applied bovine tendon type I collagen scaffold s around the sensors. The collagen scaffolds were prepared by a freeze-drying method and crosslinked using NDGA or GA treatment to minimize t heir aqueous solubility and reduce their degradation in vivo With a light microscope, we confirme d that the porous sca ffolds thoroughly surrounded the working electrodes [Fig 3.3]. Both scaffolds were semitransparent in aqueous solution. GA-cross linked scaffolds appeared white, while the NDGA-crosslinked scaffolds were brow n. We also observed high swelling for both scaffolds around the sensors in aqueous solution. We reported previously  that these sponge-like matrices wit h porous inner structure could absorb water above 99%, thus allowing gl ucose to diffuse freely. Figure 3.4 shows a schem atic of a fully assembled coil-type glucose sensor with a scaffold ready for implantat ion. The newly assembled sensor is composed of a two-electr ode system with a glucose i ndicating working Pt/Ir electrode and an Ag/AgCl reference-coun ter electrode. We added a loop between the two electrode coils to avoi d micro-shorting caused by the two electrodes touching each other. A surgic al wound clip was ap plied to provide a
60 Figure 3.3. Photographs of Implant able Sensors Coated with (A) GAcrosslinked Porous Collagen Scaffold and (B) NDGA-crosslinked Porous Collagen Scaffold.
61 2 Implanted Part Skin Outside 1 4 3 5 4 2 Implanted Part Skin Outside 1 4 3 5 4 Figure 3.4. Schematic of Short Wi re Implantable Glucose Sensor. (1) Pt/Ir working electrode with scaffold; (2) Ag/AgCl reference electrode; (3) Loop to protect micro-moti on and micro-short by two electrodes contact; (4) Wires twisted together (5) Wound clip
62 suturing site during the implantation proc edure and to prevent the sensor moving out of the skin (i.e. for anchoring). 3.3.2. Long-term In vitro Evaluation of Sensors with Porous Collagen Scaffolds The long-term in vitro function of the sensors was determined by tracking their sensitivity for up to 4 weeks. Control sensors ( without scaffold), and sensors with NDGAor GA-crosslinked scaffolds were incubated in PBS at 37C for up to 4 weeks. The sensors were removed from the PBS at weekly intervals and their sensitivity was determined. The preincubation sensitivity (week 0) was measured at the beginning of the in vitro study and the percentage of sensitivity change was calculated fr om the ratio of the sensitiv ity of the sensors at given time interval to the pre-incubation sensit ivity. The sensitivity of all sensors was tested in 5 and 15 mM glucose/PBS. Figure 3.5 s hows the sensitivity change of all sensors over 4 weeks. We observed a sli ght decrease of the sensitivity of sensors with either NDGAor GA-crosslinked scaffolds, compared to the control (no scaffold) sensors after 1 week incubation. After 2 weeks, the sensitivities of all sensors increased to a level higher th an their original sens itivity, probably because of an increase in epoxy-PU memb rane permeability due to progressive membrane swelling in aqueous solution. A fter 2 weeks, the sensitivity of the control sensors, as well as sensors with either NDGAor GA-crosslinked scaffolds, steadily decreased, however, a ll sensors retained above 80% of their original sensitivity up to 4 weeks. We believe that the sensitivity decrease after
63 1234 0 20 40 60 80 100 120 140 160 % of Sensitivity ChangeIncubation Periods (weeks) Control (no scaffold) NDGA-crosslinked GA-crosslinked Figure 3.5. Long-term In vitro Sensitivity Changes of Control Sensors and Sensors with NDGAor GA-cross linked Collagen Scaffolds. Results are shown as means SD.
64 2 weeks may be caused by progressive loss of enzyme activity. Both NDGA and GA-crosslinked scaffolds were still intact around the sensors at week 4. There was no detection of any deformation or det achment of the sc affolds from the sensor membrane surface. Although the ov erall trend of the s ensitivity of the sensors with scaffolds was lower than with t he control sensors, the application of scaffolds around sensors did not critically affect the function of the sensors during the 4 week in vitro study. 3.3.3. Long-term In vivo Performance of Sensors with Porous Collagen Scaffolds In this study, 48 sensors includi ng control sensors (short/long, CS/CL), and sensors with NDGAor GA-crosslinke d scaffolds (short/long, NS, NL, GS, GL), were implanted subcutaneously in the back of 24 Spr ague-Dawley out bred rats for a period of 4 weeks. The in vivo sensitivity of every sensor was measured at week 1, 2, 3, and 4. T he pre-implantation s ensitivity of all sensors was tested using 5 mM and 15 mM glucose/PBS just before implantation. Figure 3.6 shows a photograph of the in vivo continuous glucose monitoring procedure with the anesthetized rats. A maximum of 8 sens ors were connected to two 4 channels potentiostats. The current produced by th e sensors versus time (black arrow) was displayed on two monitors. After reaching a stable signal for 1 2 hr, glucose was administered intraperitonea lly. Small amounts of blood were sampled every 7 minutes from the ra t tail and glucose level was determined using a standard portable gluc ometer (white arrow).
65 Figure 3.6. Photograph of In vivo Continuous Glucose Monitoring Procedure.
66 The percentage of sensitivity change for each sensor during the 4 week study is shown in Figure 3.7. From this figure, a few initia l observations can be made: 1) The sensitivity of all sensors dramatically decreased after implantation compared to the pre-impl antation values. This was probably due to tissue damage which occurred during surgical procedures and the subsequent host response including protein adsorption, blood clot, and the infiltration of inflammatory cells and other cells (e.g. fibroblasts) around the sensor tips . 2) As for the in vitro study, the control sensors reta ined a higher s ensitivity than the sensors with scaffolds. The s ensors with NDGA-crosslinked collagen scaffolds also had a higher sensitivity than the sensors with GA-crosslinked scaffolds. 3) The sensitivity of the shor t sensors (CS, NS, GS) appeared to be slightly greater than the sensit ivity of the long sensors. Table 3.1 shows the number of worki ng sensors (used in Figure 3.7) at given time intervals. Initially, 3 sensor s did not work at week 1 but regained their function at week 2. Both CS and CL sens ors had a higher sensor survival rate (4 out of 8, 6 out of 8, respectively) in vivo 4 weeks post implantation than sensors with scaffold coatings. The use of scaffol ds worsened the survival rate of the sensors. However, the short sensors had a higher survival rate than the long sensors at 4 weeks post implantation (CL-4, NL-2, GL-1 vs CS-6, NS-4, GS-4). We believe that this might result from the long sensors having greater range of motion when the animals move than the short sensors.
67 01234 0 20 40 60 80 100 pre-implantationImplantation Periods (weeks)% of Sensitivity Change CS CL NS NL GS GL Figure 3.7. Long-term In vivo Sensitivity Changes of Control Sensors and Sensors with NDGAor GA-crosslinked Scaffold. Results are shown as means SD. (control s hort; CS, control long; CL, NDGAcrosslinked scaffold around short sensors; NS, NDGA-crosslinked scaffold around long sensors; NL, GA-crosslinked scaffold around short sensors; GS, GA-crosslink ed scaffold around long sensors; GL)
68 Table 3.1. Number of Working Sensor s after Implantation. (# out of 8). Weeks after implantation Scaffolds Wire 1 2 3 4 Long 8 7 4 4 Control (no scaffold) Short 7 8 7 6 Long 2 3 2 2 NDGA-crosslinked Short 8 5 5 4 Long 2 2 1 1 GA-crosslinked Shot 5 6 6 4
69 The larger macro-/micro-moti on may have caused more tissue and scaffold damage. We observed scaffold det ached from the working electrode of the long NL sensors [Fig. 3.8(A)], while the NDGA-crosslinked scaffold around the short sensor NS remained in a stabl e position [Fig. 3.8( B)]. Regarding the GA-crosslinked scaffolds, we could not detect any such scaffold around both long and short sensors [Fig. 3.8(C) and (D)]. This is consistent with our previous study where we observed that t he size and shape of the GA-crosslinked scaffolds were dramatically changed (degraded) after 4 wee ks of implantation, while the NDGAcrosslinked scaffolds remained mostly intact. In order to evaluate inflammatory response of the tissue around and within the collagen scaffolds, we direct ly implanted NDGAand GA-crosslinked scaffolds (without sensors) in the rats for up to 4 weeks. After 2 weeks implantation, H&E staining revealed the presence of many inflammatory cells including polymorphonuclear (PMN) ce lls, monocytes, and macrophages within and around the GA-crosslinked scaffolds [Fig 3.9(A)]. However, for the NDGAcrosslinked scaffolds, few inflammatory cells were observed around the scaffolds, and there was no infiltration of cells in the center region of the scaffolds [Fig. 3.9(B)]. Week 4 showed inf iltration of inflammatory ce lls and fibroblasts, along with granulation tissue deposition inside t he pore of the GA-crosslinked scaffolds [Fig. 3.9(C)], and again, less infla mmation within and around the NDGAcrosslinked scaffolds [Fig. 3.9(D)]. This result shows that the NDGA-crosslinked collagen scaffolds are more biocompatible than the GA-crosslinked collagen scaffolds and is consistent wit h a report by Koob et al.  showing that NDGA-
70 Detached Stable N.D. N.D. GA (Long) A B C D Detached Stable N.D. N.D. GA (Long) A B C D Figure 3.8. Representative Photograph of Scaffolds In situ after 4 Weeks Post Implantation. (A) Long sensor with NDGA-crosslinked scaffold; (B) Short sensor with NDGA-crosslinke d scaffold; (C) Long sensor with GA-crosslinked scaffold; (D) Short sensor with GA-crosslinked scaffold.
71 SC T B A C D T SC SC SC T T BV BV BV BV SC T B A C D T SC SC SC T T BV BV BV BV Figure 3.9. Hematoxylin and Eosin Stained Sections Showing Tissue Surrounding Porous Scaffolds. (A) GAand (B) NDGA-crosslinked scaffold and after 2 weeks post implantation; (C) GA-and (D) NDGA-crosslinked scaffold after 4 weeks post implantation. (T tissue surrounding scaffold, SC scaffold, BV blood vessels)
72 crosslinked collagen fibers appeared intact with little foreign body response after implantation in rabbits. The size of the GA-crosslinked scaffolds was reduced and the pore structure was def ormed as the implantation time increased. We also found neovasculature in both scaffolds after 4 weeks post-implantation [Fig. 3.9(C) and (D), arrows].
73 3.4. Conclusions In this study, we applied porous type I collagen scaffolds around implantable glucose sensors by a freez e-drying method and then crosslinked the scaffolds using NDGA or GA treatments. The fabricated collagen scaffolds had an open cell and interconnected pore structur e. All sensors including control sensors (without scaffold), and sensors with NDGAor GA-crosslinked scaffolds remained functional during the 4 week in vitro study. The application of both types of scaffolds around the sensors did not critically affect the function of these sensors in vitro In the 4 weeks in vivo study, the sensitivity of all sensors dramatically decreased (30 60%) after 1 week of impl antation and then remained relatively stable. The sensitivity and survival rate of the short sensors were higher than the sensitivity of the long sens ors possibly as a result of reduced motion within the animals. The sensors with NDGA-crosslinked scaffolds had a higher survival and sensitivity than the sensors with GA-cro sslinked scaffolds. By histological examination, we confi rmed that the NDGA-cross linked scaffolds are more biocompatible than the GA-crosslinked scaffolds. Therefore, this study shows that an NDGA-crosslinked collagen scaffold can be incorporated into the design of our implantable glucose sensor. However, the control sensors (no scaffolds) performed better than the sensors with scaffolds. The scaffolds alone did not impr ove the function and lifetime of our implantable glucose sensor. This indicates that in order to use these scaffolds as a way to control the local tissue envir onment around implanted sensors and thus
74 improve their function and lifetime we st ill need to improve the scaffolds. This could potentially be achieved by using t he NDGA-crosslinked collagen scaffold to also deliver various drugs and growth factors to modify the tissue response to the sensors.
75 CHAPTER 4 DEXAMETHASONE-LOADED PL GA MICROSPHERES/COLLAGEN SCAFFOLD COMPOSITE SYSTEM FOR IMPLANTABLE GLUCOSE SENSORS 4.1. Introduction Although miniaturized implantable glucose sensors show excellent performance in vitro they tend to become unreliable and lose their function after prolonged exposure to the in vivo environment, due to the foreign body response (i.e. inflammation, fibrosis, and loss of va sculature) [20,21,48,62, 98]. In particular, the accumulation of inflammatory ce lls and dense fibrotic tissue around the sensor hampers the diffusion of glucose from the capillaries to the sensors [103,113-115]. Despite numerous studies using sensors of several different types, there are no long-term implantable gluc ose sensors commercially available [62,116]. In order to improve the function of implantable glucose sensors, dexamethasone (Dex, an anti-inflammatory agent) has been used to control the tissue reactions to implanted devices. Dex, a synthetic glucocorticoid, is widely used to suppress inflammatory reacti ons caused from radiant, mechanical, chemical, infectious and immunological st imuli [50,110,117,118]. It inhibits the
76 production of critical fact ors involved in the inflammatory response such as vasoactive/ chemoattractive factors and lipol ytic/proteolytic enzymes . Patil et al.  prepared De x-loaded poly(lactic-co-glycolic) acid (PLGA) microspheres/ poly(vinyl alcohol) (PVA) hydrogel compos ite coatings for implantable biosensors to control detrimental tissue reactions and fibrosis at the sensor/tissue interface. Norton et al. [52,110] r eported that they fabric ated hydrogel (copolymer; 2hydroxy-ethyly methacrylate, N-viny l pyrrolidinon, and polyethylene glycol) sensor coatings containing Dex and/or vascular endothelia l growth factor (VEGF) to minimize the foreign body response and to promote angiogene sis. Klueh et al. [62,63] induced significant neovasculari zation surrounding an implanted sensor using a VEGF-cell-fibrin gene transfer syste m. Kim and Martin  investigated a composite of Dex-loade d PLGA nanoparticles/algi nate hydrogel for neural prosthetic application. Lincoff et al. [ 47] developed a Dex el uting stent using a high molecular weight poly-L-lactic acid (PLLA) biodegradable polymer containing the drug to prev ent restenosis. Gomez-Gaet e et al.  optimized the encapsulation of Dex in PLGA nanoparticles for ocular delivery. The use of Dex-loaded microspheres /nanospheres to provide controlled local drug delivery are typically prepared using a synthetic biodegradable polymers such as PLGA and PLLA . The degradation rate of these polymers in vitro / in vivo can be controlled by regulating t he composition of monomer units (i.e. lactic acid and glycolic acid). Thus, PLGA microspheres, which have controllable drug release kinetics, have been utilized not only for Dex delivery but also for angiogenic growth factors and other proteins delivery [61,121-123]. Both
77 PLGA and PLLA are also widely used fo r tissue engineering [124,125] and gene therapy  research due to thei r good biocompatibility and suitable biodegradability characteristics [124-126]. Due to good swelling and viscoelas tic properties and outstanding biocompatibility, hydrogels have also been used as sustained-release drug delivery systems and as the outermost coat ings of implantable glucose sensors. Pluronics [127-129], also called Poloxame rs, are particularly interesting because they are in a sol state below a lower crit ical solution temper ature (LCST; i.e., reverse sol-gel transition temperature, 420C), but transition to a gel state above 37C [130,131]. Oh et al.  fabricated a temperature-controllable crosslinked Pluronic/alginate mixture for use in de livering a non-steroidal anti-inflammatory drug (NSAID) for prevention of post-surgical tissue adhesion. In this study, we first fabricat ed porous collagen scaffolds around implantable glucose sensors using a freeze-drying method, followed by crosslinking the collagen scaffold using NDGA treatment . In order to minimize the inflammatory response to the sensors, we then added Dex-loaded microspheres to the scaffold by di pping the sensor/scaffold in a microspheres/Pluronic F127 hydrogel suspen sion. We characterized the sensors with the Dex-loaded mi crospheres/ scaffold composite system in vitro and then tested these sensors in rats.
78 4.2. Materials and Methods 4.2.1. Materials Poly(DL-lactide-co-glycolide) (P LGA, Resomer RG503H, 50:50) was a generous gifted from Boehrin ger-Ingelheim (Germany). Type I collagen (purified from fetal bovine tendon) was a generous gift from Dr. Thomas Koob, Shriners Hospital for Children (Tampa, FL). No rdihydroguaiaretic acid (NDGA) was purchased from Cayman Chemical Co. (Ann Arbor, MI). Methylene chloride (HPLC-GC/MS grade), acetonitrile (HPL C grade), methanol (HPLC grade), glucose, bovine serum albumin (BSA) and 50% (w/w) glutaraldehyde (GA) were obtained from Fisher Scientif ic (Pittsburgh, PA). Polyviny l alcohol (PVA; avg. mol. wt = 30,000 70,000), dexamethasone (Dex, C22H29FO5; Fw = 392.5), Pluronic F-127, glucose oxidase (GOD ) (EC 184.108.40.206., type X-S, Aspergillus niger 157,500 U/g), epoxy adhesive (ATACS 5104), polyurethane (PU), acetone, and tetrahydrofuran (THF) were obtained fr om Sigma-Aldrich (St. Louis, MO). Dextrose injection solution (50%, w/v) was obtained from Abbott Laboratories (North Chicago, IL). The FreeStyleTM glucometer was from TheraSense (Alameda, CA). Sprague-Dawley out-br ed rats (male, 375-399 g) were purchased from Harlan (Dublin, VA). 4.2.2. Preparation of Dex-loaded Microspheres Biodegradable PLGA microspheres l oaded with Dex were prepared by an oil-in-water (O/W) emulsion/solvent ev aporation technique. The oil phase consisted of 80 mg of PLGA and 50 mg of Dex dissolved in 6 mL of a mixture of
79 either 5:1 methylene chloride to methanol or 5:1 methylene chloride to acetone. This oil phase was added to 100 mL of 0.2% PVA in water, which was stirred with an overhead stirrer at 800 RPM for 30 min. to achieve an O/W emulsion system. The resulting emulsion was stirred on a magnetic stir plate for 16 h to allow complete evaporation of the solvent and solidification of the droplets into microspheres. During the emulsion and solidif ication process, aluminum foil was completely surrounded the beaker to prot ect from UV light (UV light will degrade Dex). The microspheres were collected by centrifugation at 8,000 RPM (7,500x g) for 15 min. in a refri gerated centrifuge set at 15 C. The microspheres were washed 5 times with deionized water. T he centrifuge tubes were capped and placed in freezer (-20C) overnight. The tubes were covered with aluminum foil and were placed in a Freeze-drying system overnight to obtain dry microspheres. 4.2.3. Microsphere Analysis The Dex loading efficiency and encapsul ation efficiency into microspheres were determined using high performance liquid chromatography (HPLC) (LC10AT vp; SPD-10A vp; SCL-10A vp; Shim adzu, Japan). Microspheres (10 mg) were dissolved in 1 mL of acetonitrile. Dex concentration in dissolved samples was determined by HPLC analysis at 246 nm using a Premier C-18 column (Shimadzu, Japan) with a m obile phase of acetonitrile and water mixture (42:58), with flowing mobile p hase at 1 mL/min.
80 The drug loading efficiency was calcul ated using the following equation: Loading efficiency (%) = mg of Dex / 10 mg of microspheres x 100 The drug encapsulation efficiency was determined using the following equation : Encapsulation efficiency (%) = experimental drug loading / theoretical drug loading x 100 The morphology of the microspheres was examined using scanning electron microscopy (SEM) after gold sputter coating of the samples in a metal evaporator according to standard procedures. 4.2.4. Preparation of Dex-loaded Mi crospheres/Scaffold Composite System Collagen scaffolds were prepared by a freeze-drying method. Collagen was dissolved in 3% acetic acid to prepar e a 1% (w/v) solution. The solution was applied to a cylinder-shaped polypropylene mold ( 10 mm, height 8 mm) and then freeze-dried. The scaffolds were then crosslinked with NDGA treatment as follow. Dried collagen scaffolds were briefl y soaked in absolute ethanol, followed by soaking in 2 M NaCl in water for 12 h at room temperature. Scaffolds were resuspended in oxygen sparged phosphate buffered saline (PBS, 0.1 M NaH2PO4, pH 9.0) for 30 min. at room temperature. Scaffo lds were then treated
81 with 3 mg NDGA/mL of PBS prepared as fo llows: NDGA was dissolved in 0.4 N NaOH at a concentration of 30 mg/mL. On e milliliter of the NDGA solution was added directly to 9 mL of PBS contai ning the scaffold. The scaffolds were agitated in the NDGA solution for 24 h at room temperature. The scaffolds were removed, briefly rinsed with water and freeze-dried. The microspheres containing Dex we re incorporated into the NDGAcrosslinked collagen scaffold by dipp ing the scaffolds in a microsphere suspensions. Two different microspher e suspension solution (hydrogel and water) were used for fabrication of mi crosphere/scaffold composites. Pluronic F127 was adopted as the hydrogel material using a 25% solution with selfaggregation properties at low critical solution temperat ure (LCST; i.e., reverse sol-gel transition temperat ure, 4-20C) [130-132]. Pl uronic solution, freshly prepared by dissolving in deionized water, was kept in the refriger ator (4C ) prior to use. Five, 10, 20 and 40 mg/mL of microspheres, loaded with Dex, were dispersed in the Pluronic solution. Dried scaffolds were soaked in the hydrogel suspension with vortex mixing to incorpor ate the microspheres evenly. During the procedure, the suspensions were kept in an ice bath to prev ent the gelation of Pluronic F-127. After completion of the loading procedure, the microspheres/scaffolds were taken out of the microspheres-hydrogel suspension solution and then placed at room temperat ure to allow gel formation. Another group of scaffolds were prepared using 5, 10, 20 and 40 mg/mL of microsphereswater suspension at room temperature. In this case, no hydrogel was used. The loading efficiency of Dex in the sca ffolds was determined using HPLC as
82 previously described above. Drug concentra tion was standardized by dividing by total mass of the scaffold including microspheres. 4.2.5. In vitro Release of Dex from the Microspheres/Scaffold Composite System The in vitro release study was performed in phosphate buffered saline (PBS) under sink conditions. Sample s (1.5 2.5 mg) of Dex-loaded microspheres/scaffold composites were incubated in 1 mL of PBS on a heated rocker (Barnstead Lab-Line, US ) at a constant temperat ure (37C) over 21 days. For comparison, 10 mg of standard De x-loaded PLGA microspheres were incubated under the same condit ions. At 3 or 7 day time intervals, 0.5 mL of supernatant was taken for analysis and repl aced with 0.5 mL of fresh PBS into the test tube. Dex concentra tion of in the samples wa s determined by HPLC, as described above. 4.2.6. Preparation of Implantable Glucose Sens ors with Microspheres/ Scaffold Composite System We first prepared coil type glucose sensors loaded with crosslinked enzyme (GOD: glucose oxidase) using a platinum-iridium (Pt/Ir) wire (Teflon coated, 0.125 mm, Pt:Ir = 9:1, Medwire, Sigmund Cohn Corp.). Bovine tendon type I collagen scaffolds were fabricated around the sensors and crosslinked with NDGA as previously described .
83 In order to incorporat e microspheres loaded with Dex into the scaffolds, the sensors with scaffolds were soaked in either the microsphere-hydrogel suspension or the microsphere-water sus pension with vortex mixing as described above. The sensitivities of the sensor s (with collagen scaffold only; and with collagen scaffold plus microspheres) were determined in 5 mM and 15 mM glucose/PBS using an Apollo 4000 potentio stat (World Precision Instruments, Inc., Sarasota, FL). Amperometric m easurements were per formed at room temperature at 0.7 V vs Ag/AgCl. The response sensitivity (S) was assessed by 1) measuring the response current (I1) of a glucose solution (C1), 2) adding a concentrated glucose solution into the m easured solution to increase the glucose concentration (C2), and 3) measuring the response current (I2) of the resulting solution. The sensitivity was expressed as the current increase caused by a 1 mM glucose increase, i.e. S = (I2 I1) / (C2-C1). 4.2.7. Implantation Procedures All implantation protocols were appr oved by the University of South Florida Institutional Animal Care and Use Committee (IA CUC). All implantable glucose sensors were prepar ed aseptically and then placed in sterile Petri dishes under humidified conditions to prevent the hydrogel fr om drying. During the surgical procedure, a continuous flow gas anesthesia system was used to deliver 1.5 % isoflurane to the rats in a 2.0 L/min. oxygen flow. Eight sensors (with microspheres/hydrogel/NDG A-crosslinked collagen scaffold) were implanted subcutaneously on the back of the rats as follows. Two
84 sensors were implanted per rat. Two 1. 5 cm long longitudinal incisions were made 1.5 cm laterally to the dorsal midl ine and 3-4 cm caudally from the neck. A subcutaneous pocket was created using blunt surgical scissors before implantation. The sensors were direct ly implanted through the incision. The sensors were secured to the skin and the incision was closed us ing 3-0 Prolene. In addition, in order to evaluate the inflammatory response to the collagen scaffolds without the influence of the sensor we directly implanted microspheres/ scaffold samples (without sensors) in t he rats. At set time intervals, tissue samples including scaffolds were embedded in paraffin. Sections (ca. 5 m in thickness) were cut and stained with Maye rs hematoxylin and eosin (H&E) stain. Stained sections were analyzed and photographed using an Olympus BX41 microscope (Olympus; Tokyo, Japan). 4.2.8. In vivo Evaluation of Sensors Coated with Microspheres/Scaffold Composite System The sensitivity of implanted sensor s was measured every seven days for up to 28 days or until there was no amper ometric response from the implanted sensors. During each measurement period, four rats were anesthetized using isofluorane and the eight im planted sensors were continuously monitored using two Apollo 4000 potentiostats. After a stable signal was obtained from the sensors, 0.7 mL of sterile 50% dextro se was administered intraperitoneally using a 27 ga. needle. Following the inject ion, small blood samples (~ 5 L) were collected every 7 min. from the rat tail, app lied to test strips and the glucose level
85 was determined using the standard FreestyleTM glucometer. The amperometric response corresponding to the glycemia of the rat was recorded at the corresponding current-time intervals of eac h sensor. The sensor sensitivity was calculated by dividing the change in curr ent (I) by the change in glycemia (C) between the initial (before dextrose inject ion) and the peak st atus (after dextrose injection) as follows: Sensitivity (nA/mM) = (Imax I0) / (Cmax C0)
86 4.3. Results and Discussion 4.3.1. Preparation of Dexloaded PLGA Microspheres In order to control the release ki netics of Dex, microspheres were fabricated using an oil-water emulsion process. We used PLGA (lactic and glycolic copolymer ratio 50:50) as t he biodegradable polymer. The microspheres had a regular spherical morphology as show n in Fig. 4.1. Th e diameter of Dexloaded microspheres varied from 1.5 to 50 m and the average diameter size was 16.0 2.3 m as estimated from SEM images in three different areas. Figure 4.1(A) shows that many De x crystals were present around the microspheres because an excess amount of Dex (50 mg in 80 mg of PLGA) was used in the microsphere preparation to in crease the Dex loading efficiency. Dex has poor solubility in water, but is fr eely soluble in alcohols. Washing the microspheres with methanol removed t he Dex crystals [Fig. 4.1(B)]. The effect of the organic solvents on Dex encapsulation was investigated using two different organic solvent system s. Methylene chloride (MC) is widely used as an organic solvent for PLGA Acetone and methanol are good solvents for Dex. A constant ratio of PLGA (80 mg ) to Dex (50 mg) and 5 mL of MC were used in this study. Table 4.1 shows De x loading efficiency and encapsulation efficiency with different solvent syste ms. The amount of Dex loading and encapsulation efficiency dramatically incr eased to 14.9 0.51 and 38.9 1.32 %, respectively, when using acetone:MC (1:5 ), compared to methanol:MC (1:5) (3.3 0.24 and 8.5 0.64 %). Bec ause acetone is also a good solvent for PLGA, the Dex-acetone solution is mo re miscible with the MC-polymer solution and thus,
87 B A B A Figure 4.1. SEM Morphology of the De x-loaded PLGA Micr ospheres. (A) with Dex crystals; (B) without Dex crystals after washing with methanol.
88 Table 4.1. Solvent Effect on th e Amount of Dex Loading Efficiency and Encapsulation Efficiency. Solvent (v:v) % of Dex Loading (/MS)Encapsulation Efficiency (%) Me-OH : MC (1:5) 3.3 0.24 8.5 0.64 Acetone : MC (1:5) 14.9 0.51 38.9 1.32
89 B A B A Figure 4.2. SEM Morphology of the De x-loaded PLGA Microspheres/Collagen Scaffold Composite. (A) x500 magnification; (B) x1,000 magnifications.
90 the amount of Dex encapsulated in PLGA was higher in the acetone/MC cosolvent system than with t he methanol/MC system due to a better continuous phase. 4.3.2. Preparation of Dex-loaded Mi crospheres/Scaffold Composite System To further in an effort suppress infl ammatory response to implantable glucose sensors, Dex-l oaded microspheres were inco rporated into the porous NDGA-crosslinked collagen scaffolds. Microspheres suspension in either Pluronic F127 hydrogel or water wa s used for the fabrication of microspheres/scaffold composites. We chose to add the microspheres to NDGAcrosslinked scaffolds to avoid Dex lo ss that would have resulted from the crosslinking method in ethanol. Figure 4.2( A) shows that the microspheres were uniformly distributed throughout the sc affold due to its open pores (with diameters ranging from 20 to 100 m) and high interconnectivity between the pores. In a higher magnification image [Fig. 4.2(B)], the Dex-loaded microspheres (1.5 50 m) can be seen attached to the collagen scaffold matrix. The effect of different suspensions on drug loading was evaluated. Figure 4.3 shows that the amount of Dex loading efficiency wa s directly proportional to the initial microsphere-loadi ng amount (5 to 20 mg/m L). Interestingly, the microspheres/scaffold composite fabricat ed using hydrogel suspension had a much higher loading effi ciency than the composite fabricated using water suspension. The Pluronic so lution (20% concentrati on) being highly viscous
91 51020 0 100 200 300 400 500 Amount of Dex Loading (g/mg of Scffold)Microspheres Content in Suspension (mg/mL) Water suspension Hydrogel suspension Figure 4.3. The Amount of Dex Loading in the Composite as Fabricated Using Either Water or Hydrogel Suspension with Different Initial Microspheres Loading Am ounts. Results are shown as mean SD (n = 4).
92 allowed the addition of many microspheres to the sc affolds during the loading process. In addition, Figure 4.4 shows t hat the amount of Dex loading decreased from 566.5 1.1 to 510.4 9. 0 g/mg of scaffold after rinsing with water for the microspheres loaded with water suspension. However, there was no significant Dex loss after water rinsing in the composite fabricat ed using hydrogel suspension. The Pluronic/microspheres mixt ures were sol state in the ice bath (below 4C), but they were gelled at room temperature after the completion of the loading process. We assume that all microspheres still remained in position inside the scaffold after the rinsing step due to this sol-gel transition behavior of the Pluronic suspension. 4.3.3. In vitro Drug Release Studies Four in vitro release studies were performe d in phosphate buffered saline (PBS) under sink conditions for both microspheres and microspheres/scaffold composite systems. Based on the loading efficiency results above, we chose to fabricate a composite system using the Pluronic hydrogel suspension. At 3 day or 7 day intervals, samples of the incubation medium were collected and Dex concentration in the supe rnatant was determi ned by HPLC. Figure 4.5(A) and 4.5(B) show the cumulative Dex releas e profiles from th e PLGA standard microspheres and the PLGA microspheres/collagen scaffold composites, respectively. An initial burst release ( 20 25 %) was observed within 6 7 days post incubation for both the microspheres and composite system. The initial burst release was probably due to residual Dex crystals on the surface of the
93 Water SuspensionHydrogel Suspension 0 200 400 600 800 Amount of Dex. Loading (g/mg of Scffold) Before rinsing After rising with water Figure 4.4. The Amount of Dex Loading in the Composite as Fabricated Using Either Water or Hydrogel Suspension after Rinsing with Water. Results are shown as mean SD (n = 4).
94 036912151821242730 0 20 40 60 80 100 Cumulative Dex Release (%)Drug Releasing Periods (Days) Collect samples every 3days Collect samples every 7days036912151821242730 0 20 40 60 80 100 Drug Releasing Periods (Days)Cumulative Dex. Release (%) Collect samples every 3days Collect samples every 7days A B036912151821242730 0 20 40 60 80 100 Cumulative Dex Release (%)Drug Releasing Periods (Days) Collect samples every 3days Collect samples every 7days036912151821242730 0 20 40 60 80 100 Drug Releasing Periods (Days)Cumulative Dex. Release (%) Collect samples every 3days Collect samples every 7days A B Figure 4.5. Cumulative Dex Releas ed from Standard Microspheres and Dexloaded Microspheres/Scaffold Composite During the In vitro Release Studies in PBS at 37C. Th e total amount of Dex released into the PBS as a percentage of the total amount of Dex encapsulated into the microspheres (A) and encapsulated into the scaffold composites (B) was plotted as a function of the elapsed time from the beginning of the release studie s. Results are shown as means SD (n = 4).
95 microspheres [50,53]. For the three day interval sample collection shows that the release of Dex from mi crospheres alone reached 90% release within 24 days, while the composite system released 50% of Dex during the same period. The composite system dramatically slowed the drug release compared to the standard microspheres. This result may su ggest that either collagen scaffold or the hydrogel phase in the scaffold delayed De x diffusion to the releasing media. The release profile of both micr ospheres and composite system when collected at 7 day intervals showed the sa me pattern with an initial burst release and continued zero order release pattern between day 7 and day 21, probably because of inadequate sink condition (PBS was replaced every 7 days). Nonetheless, the release study with 3 day sample collection showed sustained release of Dex from the microsphere/hydrogel/scaffold system over 1 month. 4.3.4. Implantable Glucose Sensors Covered with Microspheres/Scaffold Composite System We prepared coil-type glucose sensor s with porous collagen scaffolds as previously described . Then, Dex-l oaded microspheres were incorporated into the scaffolds surrounding the sensor s by soaking in microspheres-Pluronic suspensions. With a light microscope, we confirmed that the microspheres thoroughly surrounded the sensor surface [F ig. 4.6(A)]. In a higher magnification image [Fig. 4.6(B)], the De x-loaded microspheres were observed to be uniformly distributed inside the pore struct ure of the collagen scaffold.
96 A B A B Figure 4.6. Light Microscope Photographs of the Implantable Glucose Sensing Element with Dex-loaded Microspher es/Scaffold Composite. (A) Working electrode; (B) x100 magnifi cation of the scaffold region.
97 The effect of the microspheres on the function of sensors was investigated by varying the glucose concentration from 5 to 15 mM. The sensitivity change for each sensor before and after microspheres application with different suspensions is shown in Figure 4.7. We observed a s light decrease of the sensitivity of sensors with microspheres f abricated using either water or hydrogel suspension, compared to the sensors without microspher es. However, there was no statistical difference before and after microspheres application ( p > 0.05; Students t -test). Therefore, adding micros pheres around the sensors with scaffold did not negatively intact the func tion of the sensors. 4.3.5. In vivo Performance of Sensors with Dex-loaded Microspheres/ Scaffold Composite System Implantable glucose sensors with Dex-loaded microspheres/scaffold composite were implanted subcutaneously in the back of rats an d their sensitivity measured for up to 28 days or until th ere was no amperometric response. Because of higher Dex loading, we chos e to fabricate the composite system using the Pluronic hydrogel suspension containing 40 mg/mL of Dex-loaded microspheres. Figure 4.8 shows the percent sensitiv ity change of the sensors during the 2 week study. Th e sensitivity of the se nsors with composite was compared to our previous in vivo data results (without microspheres; control, NDGA-crosslinked scaffold, GA-crosslin ked scaffold). The sensors with the composite system reta ined above 50% of their original sensitivity at 2 weeks,
98 A B 2040 0 2 4 6 8 10 12 14 16 *Sensitivity (nA/mM)Microspheres Content in Suspension (mg/mL) w/o Microspheres with Microspheres* 2040 0 2 4 6 8 10 12 14 16 * Microspheres Content in Suspension (mg/mL)Sensitivity (nA/mM) w/o Microspheres with Microspheres A B 2040 0 2 4 6 8 10 12 14 16 *Sensitivity (nA/mM)Microspheres Content in Suspension (mg/mL) w/o Microspheres with Microspheres* 2040 0 2 4 6 8 10 12 14 16 * Microspheres Content in Suspension (mg/mL)Sensitivity (nA/mM) w/o Microspheres with Microspheres 2040 0 2 4 6 8 10 12 14 16 *Sensitivity (nA/mM)Microspheres Content in Suspension (mg/mL) w/o Microspheres with Microspheres* 2040 0 2 4 6 8 10 12 14 16 * Microspheres Content in Suspension (mg/mL)Sensitivity (nA/mM) w/o Microspheres with Microspheres Figure 4.7. Effect of Adding PLGA Microspheres in the Scaffold on Glucose Sensor Sensitivity with Diffe rent Suspensions. (A) Water suspension; (B) Pluronic F127 hy drogel suspension. Results are shown as means SD. (n = 4). *Indi cates no statistically significant differences before / after incorporat ion of microspheres (p > 0.05).
99 0 2 4 6 8 C ontrol N D GA scaffold G A s c af f o l d N D GA scaffold /M S 0 20 40 60 # of working sensorNumber of Working Sensor Implantation Periods 2 weeks% of Sensitivity Change Figure 4.8. In vivo Sensitivity Changes (Bar Gr aph results are shown as means SD) and Number of Work ing Sensors (Line Graph) of Control Sensors and Sensors wit h NDGAor GA-crosslinked Collagen Scaffolds and Sensors wit h Dex-loaded Microspheres/ NDGA-crosslinked Collagen Scaffold after 2 Weeks Post Implantation.
100 while the sensitivity of the control sensors, sensors with NDGA-crosslinked scaffolds and sensors with GA-crosslinked scaffolds decreased to 42%, 30%, and 15%, respectively. We believe that this was because the locally delivered Dex effectively decreased the inflammatory response to the sensors. However, it was observed that only 2 out of 8 sensors with the composite scaffolds functioned at 2 weeks. We suggest that th e reason for the functional failure may be related to the reference electrode [Fig 4.9(A)]. Figure 4.9(B) shows a dense fibrous capsule surrounding the reference electrode. In this study, we applied Dex-loaded microspheres/scaffold com posite around the working electrode but not around the reference electrode. For this study, we did not coil the reference electrode to avoid micro-shorting caused by touching reference electrode coil (Ag/AgCl wire) to the uncovered Pt/Ir wire. However, this different reference electrode geometry may have induced a larger inflamed area of tissue. After 4 weeks, we excised non-f unctional sensors and tested them ex vivo (with tissue) in 5 mM and 15 mM glucose/PBS. Subs equently, the sensors were removed from the surrounding tissue and tested in vitro Figure 4.10 shows the amperometric response curves of an ex planted glucose sensor with and without fibrous capsule tissue. It was found that the sensor with it s fibrous capsule responded poorly (line A, sensitivity = 1.5 nA/mM) to changes in glucose concentration, while the sens or regained its initial functi on (line B, sensitivity = 15.4 nA/mM) after removing the surrounding tissue. These shows that the dense fibrous capsule tissue which forms around the reference electrode can also affect
101 R B W R A R B W R A Figure 4.9. Light Microscope Photographs of Implantable Glucose Sensors. (A) Reference and working electrode r egion; (B) Dense fibrous capsule tissue surrounding the reference elec trode (R reference electrode, W working electrode).
102 100015002000250030003500 60 80 100 120 140 160 180 B A S = 1.5 (nA/mM) S = 15.4 (nA/mM) Current (nA)Time (sec.) Glucose Conc. 5 mM >> 15 mM Figure 4.10. Amperometric Response Curv es of the Explanted Non-functioned Glucose Sensors after 4 Wee ks Post Implantation. (A) Ex vivo response of the sensor with surrounding fibrous capsule tissue; (B) In vitro response of the sensor a fter removing su rrounding the fibrous capsule tissue, after glucose concentration increase from 5 to 15 mM.
103 the sensor. Thus, the reference elec trode could also be surrounded by the composite system described in this paper. 4.3.6. Suppression of In flammation to Dex-loaded Microspheres/Scaffold Composite System To confirm the anti-inflammator y response of the Dex-loaded microspheres/scaffold composite (40 mg /mL of microspheres, Pluronic F127 hydrogel suspension), we implant ed the composites (without sensors) subcutaneously in rats Standard NDGA-crosslinked scaffolds (without microspheres) were implanted for compar ison. The histological results (H&E stain) for sampled at 2 and 4 weeks afte r implantation for both the scaffold alone and the composite scaffold are shown in Figure 4.11. The inflammatory cells were stained as purple, while normal cells were stained as pink. A very strong inflammatory response was shown around the control scaffold 2 weeks after implantation [Fig. 4.11(A)]. Predominant polymorphonucl ear leukocytes (PMNs) with monocytes and macrophages were obs erved and a dense connective tissue layer (fibrous capsule) surrounded the peri phery of the scaffold. In contrast, the inflammatory response to the Dexloaded composites [F ig. 4.11(B)] was diminished compared to t he control scaffold. The hist ological results after one week implantation are not shown as t here was no noticeable difference between the control scaffold and the Dex-loaded composites. After 4 weeks post implantation, the inflammatory response to the Dex-released composites was low [Fig. 4.11(D)] while a severe inflammatory response with a thick fibrous capsule
104 A C D SC CT SC CT B SC CT SC CT A C D SC CT SC CT B SC CT SC CT Figure 4.11. Hematoxylin and Eosin St ained Sections of Tissue Surrounding Porous Scaffolds in Rats. (A) NDGA-crosslinked scaffold; (B) Dexloaded microspheres / NDGA-crosslin ked scaffold composite after 2 weeks post implantation; (C) NDGA-crosslinked scaffold; (D) Dex-loaded microspher es / NDGA-crosslinked scaffold composite after 4 weeks post implantation (CT connective tissue, SCscaffold).
105 was present around the scaffold without De x-loaded microspheres [Fig. 4.11(C)]. These results demonstrate that the Dex released from the microspheres/scaffold greately reduced the inflammato ry response to the scaffold.
106 4.4. Conclusions In this study, we prepared Dexloaded PLGA microspheres and incorporated then into three dim ensional porous type I collagen scaffolds (crosslinked with NDGA) around implantable glucose sensors. The fabricated composite has effectively loaded Dex and sustained release of Dex for at least one month. The composite system did not si gnificantly alter the function of the sensors in vitro despite the high amount of microspheres. After 2 weeks in vivo the sensitivity of sensors with the com posite system remained higher than for other sensors without the composite system The histological results showed that the inflammatory response was lowered using the Dex-loaded composite scaffold when compared to the inflammatory response to the scaffolds without Dexloaded microspheres at 2 and 4 weeks afte r implantation. These results showed that our Dex-load ed composite system reduces inflammation around the implanted glucose sensors ti ps and could potentially impr ove their function and lifetime.
107 CHAPTER 5 SUGGESTIONS FOR FUTURE STUDY The best tissue environment for impl antable biosensor is vascularized tissue around sensor. The control of neovascu larization has recently focused on the use of angiogenic growth factors su ch as VEGF and PDGF. Norton et al. [52,110] reported that they fabricated hydrogel sensor coatings containing Dex and/or VEGF to minimize foreign body response and to promote angiogenesis. Klueh et al. [62,63] induced signific ant neovascularization surrounding an implanted sensor using a VEGF-ce ll-fibrin gene transfer system. The goal future studies will be to in troduce local delivery with microsphere systems to release angiogenic factors (VEGF, PDGF) and anti-inflammatory drugs (i.e. dexmethasone), concurrently Since Dex can lead to an antiangiogenesis effect along with an anti-infl ammatory response, the optimization of the concentrations of either angiogenic fact ors or Dex will play an important role in the dual release system. The future investigations should determine how to control neovasculrature dens ity without foreign body response around implanted glucose sensors. In addi tion, we found a dense fibr ous capsule surrounding the reference electrode in Chapter 4. In future investigations, the reference electrode should also be coated with the same approac h utilized for the working electrode.
108 REFERENCES 1. ADA. American Diabetes Associ ation All about diabetes. URL http://www.diabetes.org/ about-diabetes.jsp 2008. 2. Newman JD, Turner AP. Home bl ood glucose biosensors: a commercial perspective. Biosens Bi oelectron 2005;20(12):2435-53. 3. Heller A. Implanted electrochemic al glucose sensors for the management of diabetes. Annu Rev Biomed Eng 1999;1:153-75. 4. NIH-NIDDK. National Institute of Diabetes and Digestiv and Kidney Diseases Diabetes overview. URL http://diabetes.niddk.nih .gov/dm/pubs/overview/Dia betesOverview.pdf 2008. 5. ADA. American Diabetes Associati on; Self-monitoring of blood glucose. Diabetes Care 1994;17(1):81-6. 6. Lee S, Nayak V, Dodds J, Pishko M, Smith NB. Glucose measurements with sensors and ultrasound. Ultras ound Med Biol 2005;31(7):971-7. 7. Moussy F, Harrison DJ, O'Brien DW, Rajotte RV. Performance of subcutaneously implanted needle-type glucose sensors employing a novel trilayer coating. Anal Chem 1993;65(15):2072-7. 8. Johnson KW, Mastrototaro JJ, Howey DC, Brunelle RL, Burden-Brady PL, Bryan NA, Andrew CC, Rowe HM, Alle n DJ, Noffke BW and others. In vivo evaluation of an electroenzym atic glucose sensor im planted in subcutaneous tissue. Biosens Bioele ctron 1992;7(10):709-14. 9. Koudelka M, Rohner-Jeanr enaud F, Terrettaz J, B obbioni-Harsch E, de Rooij NF, Jeanrenaud B. In-vivo beha viour of hypodermically implanted microfabricated glucose sensors. Biosens Bioel ectron 1991;6(1):31-6.
109 10. Bindra DS, Zhang Y, Wilson GS, St ernberg R, Thevenot DR, Moatti D, Reach G. Design and in vitro studies of a needle-type glucose sensor for subcutaneous monitoring. A nal Chem 1991;63(17):1692-6. 11. Pickup JC, Shaw GW, Claremont DJ In vivo molecular sensing in diabetes mellitus: an implant able glucose sensor with direct electron transfer. Diabetologia 1989;32(3):213-7. 12. Shichiri M, Asakawa N, Yamasaki Y, Kawamori R, Abe H. Telemetry glucose monitoring device with needle-type glucose sensor: a useful tool for blood glucose monitoring in diabetic indivi duals. Diabetes Care Volume 9; 1986. p 298-301. 13. Ertefai S, Gough DA. Physiological preparation for studying the response of subcutaneously implanted glucose and oxygen sensors. J Biomed Eng 1989;11(5):362-8. 14. Armour JC, Lucisano JY, McKean BD, Gough DA. Application of chronic intravascular blood glucose sensor in dogs. Diabetes 1990;39(12):1519-26. 15. Frost MC, Meyerhoff ME. Implantabl e chemical sensor s for real-time clinical monitoring: progress and challenges. Curr Opin Chem Biol 2002;6(5):633-41. 16. Ash SR, Poulos JT, Rainier JB Zopp WE, Janle E, Kissinger PT. Subcutaneous capillary filtrate collector for measurement of blood glucose. Asaio J 1992;38(3):M416-20. 17. Meyerhoff C, Bischof F, Sternber g F, Zier H, Pfeiffer EF. On line continuous monitoring of subcutaneous ti ssue glucose in men by combining portable glucosensor with microdialys is. Diabetologia 1 992;35(11):1087-92. 18. Moscone D, Mascini M. Microdialysi s and glucose biosensor for in vivo monitoring. Ann Biol Clin (Paris) 1992;50(5):323-7. 19. Joseph JI, Torjman MJ. Glucose S ensors. In: Wnek GE, Bowlin GL, editors. Encyclopedia of biomaterials and biomedical engineering. New York: Marcel Dekker; 2004. p 683-692.
110 20. Reichert WM, Saavedra SS. Material s considerations in the selection, performance, and adhesion of polymeric en capsulants for implantable sensors. In: Williams DF, editor. Medical and Dental Materials. New York: VCH Publishers, Inc.; 1992. p 303-343. 21. Reichert WM, Sharkawy AA. C hap. 28 Biosonsors In: Von Recum A, editor. Handbook of biomaterials evaluation : scientific, technical, and clinical testing of implant materials. Philadelph ia: Taylor & Francis; 1999. p 439-460. 22. Sharkawy AA, Neuman MR, Reicher t WM. Sensocompatibility: Design considerations for biosensor-based drug delivery systems. In: Park K, editor. Controlled Drug Delivery: The Next G eneration. Washingt on, DC: American Chemical Society; in Press. 23. Clark LC, Jr., Lyons C. Electr ode systems for continuous monitoring in cardiovascular surgery. A nn N Y Acad Sci 1962;102:29-45. 24. Updike SJ, Hicks GP. The enzyme electrode. Nature 1967;214(92):986-8. 25. Gough DA, Leypoldt JK, Armour JC Progress toward a potentially implantable, enzyme-based glucose sens or. Diabetes Care 1982;5(3):190-8. 26. Pickup JC, Claremont DJ, Shaw GW. Responses and calibration of amperometric glucose sensors implanted in the subcutaneous tissue of man. Acta Diabetol 1993;30(3):143-8. 27. Abel PU, von Woedtke T. Biosensor s for in vivo glucose measurement: can we cross the experimental stage. Biosens Bioelectron 2002;17(11-12):105970. 28. Anderson JM. Chapter 4 Mechanism s of inflammation and infection with implanted devices. 1993;2( 3, Supplement 1):33. 29. Johnston RB, Jr. Current conc epts: immunology. Monocytes and macrophages. N Engl J Med 1988;318(12):747-52. 30. Maciag T. Molecular and cellular mechanisms of angiogenesis. Important Adv Oncol 1990:85-98.
111 31. Thompson JA, Anderson KD, DiPie tro JM, Zwiebel JA, Zametta M, Anderson WF, Maciag T. Site-directed neovessel formation in vivo. Science 1988;241(4871):1349-52. 32. Ziats NP, Miller KM, Anderson JM. In vi tro and in vivo interactions of cells with biomaterials. Biomat erials 1988;9(1):5-13. 33. Sharkawy AA, Klitzm an B, Truskey GA, Reichert WM. Engineering the tissue which encapsulates subcutaneous implants. I. Diffusion properties. J Biomed Mater Res 1997;37(3):401-12. 34. Cannas M, Bosetti M, Santin M, Mazzarelli S. 2. Tissue response to implants: Molecular interactions and histol ogical correlation. In: Wise DL, editor. Biomaterials and bioengine ering handbook. New York: Marcel Dekker; 2000. p 95-117. 35. Quinn CP, Pathak CP, Heller A, Hubbell JA. Photo-crosslinked copolymers of 2-hydroxyethyl methacryla te, poly(ethylene glycol) tetra-acrylate and ethylene dimethacrylate for improv ing biocompatibility of biosensors. Biomaterials 1995;16(5):389-96. 36. Rigby GP, Crump P, Vadgama P. O pen flow microperfusion: approach to in vivo glucose monitoring. Med Biol Eng Comput 1995;33(2):231-4. 37. Reddy SM, Vadgama PM. Ion ex changer modified PVC membranes-selectivity studies and re sponse amplification of ox alate and lactate enzyme electrodes. Biosens Bioelec tron 1997;12(9-10):1003-12. 38. Shichiri M, Yamasaki Y, Nao K, Seki ya M, Ueda N. In vivo characteristics of needle-type glucose sensor--measur ements of subcutaneous glucose concentrations in human volunteers. Ho rm Metab Res Supp l 1988;20:17-20. 39. Shaw GW, Claremont DJ, Pickup JC In vitro testing of a simply constructed, highly stable glucose sensor suitable for implant ation in diabetic patients. Biosens Bioel ectron 1991;6(5):401-6. 40. Wilkins E, Atanasov P, Muggenburg BA. Integrated implantable device for long-term glucose monitoring. Bios ens Bioelectron 1995;10(5):485-94.
112 41. Moussy F, Jakeway S, Harrison DJ Rajotte RV. In vitro and in vivo performance and lifetime of perfluorinated ionomer-coated glucose sensors after high-temperature curing. A nal Chem 1994;66(22):3882-8. 42. Moussy F, Harrison DJ, Rajotte RV. A miniaturized Nafion-based glucose sensor: in vitro and in vivo evaluation in dogs. Int J Artif Organs 1994;17(2):88-94. 43. Moussy F, Harrison DJ. Prevent ion of the rapid degradation of subcutaneously implanted Ag/AgCl refer ence electrodes using polymer coatings. Anal Chem 1994;66(5):674-9. 44. Kerner W, Kiwit M, Linke B, Keck FS Zier H, Pfeiffer EF The function of a hydrogen peroxide-detecting electroenzymat ic glucose electrode is markedly impaired in human sub-cutaneous tissu e and plasma. Biosens Bioelectron 1993;8(9-10):473-82. 45. Reil TD, Sarkar R, Kashyap VS, Sarkar M, Gelabert HA. Dexamethasone suppresses vascular smooth muscle ce ll proliferation. J Surg Res 1999;85(1):109-14. 46. Kawamura M, Hatanaka K, Saito M, Ogino M, Ono T, Ogino K, Matsuo S, Harada Y. Are the anti-inflammatory ef fects of dexamethasone responsible for inhibition of the induction of enzymes in volved in prostanoid formation in rat carrageenin-induced pleurisy? Eur J Pharmacol 2000;400(1):127-35. 47. Lincoff AM, Furst JG, Ellis SG, Tuch RJ, Topol EJ. Sustained local delivery of dexamethasone by a novel intravascular el uting stent to prevent restenosis in the porcine coronary in jury model. J Am Coll Cardiol 1997;29(4):808-16. 48. Wisniewski N, Moussy F, Reichert WM. Characterization of implantable biosensor membrane biofouling. Freseni us J Anal Chem 2000;366(6-7):611-21. 49. Burgess DJ, Hickey AJ. Microsphere Technology and Applications In: Swarbrick J, Boylan JC, editors. Encyclop edia of Pharmaceutical Technology: Marcel Dekker Inc; 1994. p 1-29.
113 50. Hickey T, Kreutzer D, Burgess DJ, Moussy F. Dexamethasone/PLGA microspheres for continuous delivery of an anti-inflammatory drug for implantable medical devices. Biomat erials 2002;23(7):1649-56. 51. Hickey T, Kreutzer D, Burgess DJ, M oussy F. In vivo evaluation of a dexamethasone/PLGA microsphere system designed to suppress the inflammatory tissue response to implant able medical devices. J Biomed Mater Res 2002;61(2):180-7. 52. Norton LW, Tegnell E, Toporek SS, Re ichert WM. In vitro characterization of vascular endothelial growth factor and dexamethasone releasing hydrogels for implantable probe coatings. Bi omaterials 2005;26(16):3285-97. 53. Patil SD, Papadimitrakopoulos F, Burgess DJ. Dexamethasone-loaded poly(lactic-co-glycolic) acid microspheres /poly(vinyl alcohol) hydrogel composite coatings for inflammation control. Di abetes Technol Ther 2004;6(6):887-97. 54. Yancopoulos GD, Davis S, Gale NW, Rudge JS, Wiegand SJ, Holash J. Vascular-specific growth factors and blood vessel formation. Nature 2000;407(6801):242-8. 55. Darland DC, D'Amore PA. Blood vesse l maturation: vascular development comes of age. J Clin Invest 1999;103(2):157-8. 56. Lokmic Z, Idrizi R, Messina A, Kn ight KR, Morrison W, Mitchell G. Vascularization of Engineered Constructs In: Wnek GE, Bowlin GL, editors. Encyclopedia of biomaterials and biomed ical engineering. New York: Marcel Dekker; 2004. p 1750-1759. 57. Benjamin LE, Hemo I, Keshet E. A plasticity window for blood vessel remodelling is defined by pericyte cove rage of the preformed endothelial network and is regulated by PDGF-B and VEG F. Development 1998;125(9):1591-8. 58. Ennett AB, Mooney DJ. Tissue engine ering strategies for in vivo neovascularisation. Expert Opin Biol Ther 2002;2(8):805-18.
114 59. Prokop A, Kozlov E, Nun Non S, Dikov MM, Sephel GC, Whitsitt JS, Davidson JM. Towards retrievable vascularized bioartificial pancreas: induction and long-lasting stability of polymeric mesh implant vascularized with the help of acidic and basic fibroblast growth factor s and hydrogel coati ng. Diabetes Technol Ther 2001;3(2):245-61. 60. Chen RR, Mooney DJ. Polymeric grow th factor delivery strategies for tissue engineering. Pharm Res 2003;20(8):1103-12. 61. Richardson TP, Peters MC, Ennett AB, Mooney DJ. Polymeric system for dual growth factor delivery. Nat Biotechnol 2001;19(11):1029-34. 62. Klueh U, Dorsky DI, Kr eutzer DL. Use of vascular endothelial cell growth factor gene transfer to enhance implantable sensor function in vivo. J Biomed Mater Res A 2003;67(4):1072-86. 63. Klueh U, Dorsky DI, Kreutzer DL. Enhancement of implantable glucose sensor function in vivo using gene transfer-induced neovascularization. Biomaterials 2005;26(10):1155-63. 64. Ward WK, Quinn MJ, Wood MD, Tieko tter KL, Pidikiti S, Gallagher JA. Vascularizing the tissue surrounding a model biosensor: how localized is the effect of a subcutaneous infusion of vascular endothelial growth factor (VEGF)? Biosens Bioelectron 2003;19(3):155-63. 65. Sheu MT, Huang JC, Yeh GC, Ho HO. Characterization of collagen gel solutions and collagen matrices for cell cu lture. Biomaterials 2001;22(13):1713-9. 66. Pieper JS, van der Kr aan PM, Hafmans T, Kamp J, Buma P, van Susante JL, van den Berg WB, Veerkamp JH, v an Kuppevelt TH. Crosslinked type II collagen matrices: preparatio n, characterization, and potential for cartilage engineering. Biomateria ls 2002;23(15):3183-92. 67. Chvapil M, Kronenthal L, Van Wi nkle W, Jr. Medical and surgical applications of collagen. Int Rev Connect Tissue Res 1973;6:1-61. 68. Pachence JM. Collagen-based devices for soft tissue repair. J Biomed Mater Res 1996;33(1):35-40.
115 69. Lee CH, Singla A, Lee Y. Biomedical applications of collagen. Int J Pharm 2001;221(1-2):1-22. 70. Ramshaw JA, Werkmeis ter JA, Glattauer V. Collagen-based biomaterials. Biotechnol Genet E ng Rev 1996;13:335-82. 71. Ramshaw JAM, Glattauer V, Werkme ister JA. Stabilization of collagen in medical devices. In: Wise DL, editor. Biomaterials and bioengineering handbook. New York: Marcel Dekker; 2000. p 717-738. 72. Sano A, Hojo T, Maeda M, Fujiok a K. Protein release from collagen matrices. Adv Drug Deliv Rev 1998;31(3):247-266. 73. Friess W. Collagen--biomaterial fo r drug delivery. Eu r J Pharm Biopharm 1998;45(2):113-36. 74. Griffith M, Hakim M, Shimmura S, Watsky MA, Li F, Carlsson D, Doillon CJ, Nakamura M, Suuronen E, Shinozaki N and others. Artificial human corneas: scaffolds for transplantation and host regeneration. Cornea 2002;21(7 Suppl):S54-61. 75. Barbani N, Giusti P, Lazzeri L, Polacco G, Pizzirani G. Bioartificial materials based on collagen: 1. Co llagen cross-linking with gaseous glutaraldehyde. J Biomater Sc i Polym Ed 1995;7(6):461-9. 76. Nimni ME, Cheung B, Strates B, K odama M, Sheikh K. Bioprosthesis derived from crosslinked and chemically modified collageneous tissue. In: Nimni ME, editor. Collagen Vol. III Biotechnology Boca Raton, FL: CRC Press; 1998. p 1-38. 77. Huang-Lee LL, Cheung DT, Nimni ME. Biochemical changes and cytotoxicity associated with the degradation of polymeric glutaraldehyde derived crosslinks. J Biomed Ma ter Res 1990;24(9):1185-201. 78. van Luyn MJ, van Wachem PB, Olde Damink LH, Dijkstra PJ, Feijen J, Nieuwenhuis P. Secondary cytotoxicity of cross-linked dermal sheep collagens during repeated exposure to human fibrobl asts. Biomaterials 1992;13(14):101724.
116 79. Khor E. Methods for the treat ment of collagenous tissues for bioprostheses. Biomater ials 1997;18(2):95-105. 80. Sung HW, Hsu HL, Shih CC, Lin DS Cross-linking characteristics of biological tissues fixed with monofunctional or multifunctional epoxy compounds. Biomaterials 1996;17(14):1405-10. 81. Chvapil M. Considerati ons on manufacturing principl es of a synthetic burn dressing: a review. J Biom ed Mater Res 1982;16(3):245-63. 82. Chvapil M, Speer D, Mora W, Eskelson C. Effe ct of tanning agent on tissue reaction to tissue implanted co llagen sponge. J Surg Res 1983;35(5):4029. 83. Lee CR, Grodzinsky AJ, Spector M. The effects of cross-linking of collagen-glycosaminoglycan scaffolds on co mpressive stiffness, chondrocytemediated contraction, proliferation and biosynthesis. Biomaterials 2001;22(23):3145-54. 84. Tsai CL, Hsu Sh SH, Cheng WL. Effect of different solvents and crosslinkers on cytocompatibility of Type II collagen scaffolds for chondrocyte seeding. Artif Organs 2002;26(1):18-26. 85. Koob TJ. Biomimetic approaches to tendon repair. Comp Biochem Physiol A Mol Integr Physiol 2002;133(4):1171-92. 86. Koob TJ, Hernandez DJ. Material properties of polymerized NDGAcollagen composite fibers: dev elopment of biologically based tendon constructs. Biomaterials 2002;23(1):203-12. 87. Koob TJ, Willis TA, Hernandez DJ. Biocompatibility of NDGA-polymerized collagen fibers. I. Evaluatio n of cytotoxicity with tendon fibroblasts in vitro. J Biomed Mater Res 2001;56(1):31-9. 88. Koob TJ, Willis TA, Qiu YS, Her nandez DJ. Biocompatibility of NDGApolymerized collagen fibers. II. Attachment proliferation, and migration of tendon fibroblasts in vitro. J Bi omed Mater Res 2001;56(1):40-8.
117 89. Harkness RD. Collagen. Sci Prog 1966;54(214):257-74. 90. Stenzel KH, Dunn MW, Rubin AL, Mi yata T. Collagen gels: design for a vitreous replacement. Science 1969;164(885):1282-3. 91. Miyata T, Sode T, Rubin AL, Stenzel KH. Effects of ultr aviolet irradiation on native and telopeptide-poor collagen. Bi ochim Biophys Acta 1971;229(3):67280. 92. Gorham SD, Light ND, Diamond AM, Willins MJ, Bailey AJ, Wess TJ, Leslie NJ. Effect of chemical modificati ons on the susceptibility of collagen to proteolysis. II. Dehydrothermal crosslinki ng. Int J Biol Macromol 1992;14(3):12938. 93. Long N, Yu B, Moussy Y, Moussy F. Strategies for testing long-term transcutaneous amperometric glucose sensors. Diabetes Technol Ther 2005;7(6):927-36. 94. Yu B, Long N, Moussy Y, Moussy F. A long-term flexible minimallyinvasive implantable glucose bios ensor based on an epoxy-enhanced polyurethane membrane. Biosen s Bioelectron 2006;21(12):2275-82. 95. Koob TJ. Collagen Fixation. In: Wnek GE, Bowlin GL, editors. Encyclopedia of biomaterials and biomed ical engineering. New York: Marcel Dekker; 2004. p 335-347. 96. Patel VR, Amiji MM. Preparation and characterization of freeze-dried chitosan-poly(ethylene oxide) hydrogels for site-specific antibiotic delivery in the stomach. Pharm Re s 1996;13(4):588-93. 97. Angele P, Abke J, Kujat R, Falte rmeier H, Schumann D, Nerlich M, Kinner B, Englert C, Ruszczak Z, Mehrl R and others. Influenc e of different collagen species on physico-chemical properties of crosslinked collagen matrices. Biomaterials 2004;25(14):2831-41. 98. Mercado RC, Moussy F. In vitro and in vivo mineralization of Nafion membrane used for implantable glucos e sensors. Biosens Bioelectron 1998;13(2):133-45.
118 99. Sharkawy AA, Klitzm an B, Truskey GA, Reichert WM. Engineering the tissue which encapsulates subcutaneous im plants. II. Plasma-tissue exchange properties. J Biomed Mate r Res 1998;40(4):586-97. 100. Sharkawy AA, Klitzm an B, Truskey GA, Reichert WM. Engineering the tissue which encapsulates subcutaneous impl ants. III. Effective tissue response times. J Biomed Mater Res 1998;40(4):598-605. 101. Pfeiffer EF. On the way to the aut omated (blood) glucose regulation in diabetes: the dark past, the grey present and the rosy future. XII Congress of the International Diabetes Federation, M adrid, 22-28 September 1985. Diabetologia 1987;30(2):51-65. 102. Clark LC, Jr., Noyes LK, Spokane RB, Sudan R, Miller ML. Long-term implantation of voltammetr ic oxidase/peroxide gluc ose sensors in the rat peritoneum. Methods Enzym ol 1988;137:68-89. 103. Updike SJ, Shults MC, Rhodes RK, Gilligan BJ, Luebow JO, von Heimburg D. Enzymatic glucose sensor s. Improved long-term performance in vitro and in vivo. Asaio J 1994;40(2):157-63. 104. Gilligan BJ, Shults MC, Rhodes RK, Updike SJ. Evaluation of a subcutaneous glucose sensor out to 3 months in a dog model. Diabetes Care 1994;17(8):882-7. 105. Nishida K, Sakakida M, Ichinose K, Uemura T, Uehara M, Kajiwara K, Miyata T, Shichiri M, Ishihara K, Nak abayashi N. Developm ent of a ferrocenemediated needle-type glucose sens or covered with newly designed biocompatible membrane, 2-methacryloyl oxyethyl phosphorylcholine-co-n-butyl methacrylate. Med Prog Technol 1995;21(2):91-103. 106. Quinn CA, Connor RE, Heller A. Biocompatible, glucose-permeable hydrogel for in situ coating of im plantable biosensor s. Biomaterials 1997;18(24):1665-70. 107. Lei M, Baldi A, Nuxoll E, Siegel RA, Ziaie B. A hydr ogel-based implantable micromachined transponder for wirele ss glucose measurement. Diabetes Technol Ther 2006;8(1):112-22.
119 108. Fernandez E, Lopez-Cabarcos E, Mijangos C. Viscoelastic and swelling properties of glucoe oxidas e loaded polyacrylamide hy drogels and the evaluation of their properties as glucose sensors. Polyme r 2005;46(7):2211-2217. 109. Gifford R, Batchelor MM, Lee Y, Gokulrangan G, Meyerhoff ME, Wilson GS. Mediation of in vivo glucose sensor inflammatory response via nitric oxide release. J Biomed Mater Res A 2005;75(4):755-66. 110. Norton LW, Koschwanez HE, Wisnie wski NA, Klitzman B, Reichert WM. Vascular endothelial growth factor and dexamethasone release from nonfouling sensor coatings affect the foreign body response. J Biomed Mater Res A 2007;81(4):858-69. 111. Ju YM, Yu B, Koob TJ, Moussy Y, Moussy F. A novel porous collagen scaffold around an implantable biosensor for improving biocompat ibility. I. In vitro/in vivo stability of the scaffold and in vitro sensitivity of the glucose sensor with scaffold. J Biomed Mater Res 2007;in press. 112. Dungel P, Long N, Yu B, Moussy Y, Moussy F. Study of the effects of tissue reactions on the function of impl anted glucose sensors. J Biomed Mater Res A 2007. 113. Wilson GS, Zhang Y, Reach G, Moatti -Sirat D, Poitout V, Thevenot DR, Lemonnier F, Klein JC. Progress toward the development of an implantable sensor for glucose. Clin Chem 1992;38(9):1613-7. 114. Updike SJ, Shults MC, Gilligan BJ, Rhodes RK. A subcutaneous glucose sensor with improved longev ity, dynamic range, and st ability of calibration. Diabetes Care 2000;23(2):208-14. 115. Moatti-Sirat D, Capron F, Poitout V, Reach G, Bindra DS, Zhang Y, Wilson GS, Thevenot DR. Towards continuous glucose monitoring: in vivo evaluation of a miniaturized glucose sensor implanted for several days in rat subcutaneous tissue. Diabetologia 1992;35(3):224-30. 116. Kissinger PT. Biosensors-a perspective. Biosens Bioelectron 2005;20(12):2512-6.
120 117. Yoon JJ, Kim JH, Park TG. Dexa methasone-releasing biodegradable polymer scaffolds fabricated by a gas-foam ing/salt-leaching method. Biomaterials 2003;24(13):2323-9. 118. Schleimer R, Busse W, O'Brien P. Inhaled glucocorticoids in asthma. In: Swarbrick J, Boylon JC, editors. Encycl opedia of pharmaceutical technology 1997;New York, Basel: Maecel Dekker:1-29. 119. Kim D-H, Martin DC. Sustained release of dexamethasone from hydrophilic matrices using PLGA nano particles for neural drug delivery. Biomaterials 2006;27:3031-3037. 120. Gomez-Gaete C, Tsapis N, Besnard M, Bochot A, Fattal E. Encapsulation of dexamethasone into biodegradable pol ymeric nanoparticles. Int J Pharm 2007;331(2):153-9. 121. Sanders LM, Kell BA, McRae G I, Whitehead GW. Prolonged controlledrelease of nafarelin, a luteinizing hormone-releasing hormone analogue, from biodegradable polymeric impl ants: influence of composit ion and molecular weight of polymer. J Pharm Sci 1986;75(4):356-60. 122. Heya T, Okada H, Ogawa Y, Toguchi H. In vitro and in vivo evaluation of thyrotrophin releasing hormone release from copoly(dl-lactic/glycolic acid) microspheres. J Pharm Sci 1994;83(5):636-40. 123. Bourke SL, Al-Khalili M, Briggs T, Michniak BB, Kohn J, Poole-Warren LA. A photo-crosslinked poly(vinyl alcohol) hydr ogel growth factor release vehicle for wound healing applications. AA PS PharmSci 2003;5(4):E33. 124. Wu YC, Shaw SY, Lin HR, Lee TM Yang CY. Bone tissue engineering evaluation based on rat calvaria stro mal cells cultured on modified PLGA scaffolds. Biomateria ls 2006;27(6):896-904. 125. Mikos AG, Sarakinos G, Leite SM Vacanti JP, Langer R. Laminated three-dimensional biodegr adable foams for use in tissue engineering. Biomaterials 1993;14(5):323-30.
121 126. Richardson TP, Murphy WL, Mooney DJ Polymeric delivery of proteins and plasmid DNA for tissue engineering and gene therapy. Crit Rev Eukaryot Gene Expr 2001;11( 1-3):47-58. 127. Leach RE, Henry RL. Reduction of postoperative adhesions in the rat uterine horn model with poloxamer 407. Am J Obstet Gynec ol 1990;162(5):13179. 128. Steinleitner A, Lambert H, Kazensky C, Cantor B. Poloxamer 407 as an intraperitoneal barrier material for t he prevention of postsurgical adhesion formation and reformation in rodent models for reproductive surgery. Obstet Gynecol 1991;77(1):48-52. 129. Steinleitner A, Lopez G, Suarez M, Lambert H. An evaluation of Flowgel as an intraperitoneal barrier for prevention of postsur gical adhesion reformation. Fertil Steril 1992; 57(2):305-8. 130. Yamaoka T, Takahashi Y, Fujisato T, Lee CW, Tsuji T, Ohta T, Murakami A, Kimura Y. Novel adhesion prevent ion membrane based on a bioresorbable copoly(ester-ether) comprised of poly-L-lact ide and Pluronic: in vitro and in vivo evaluations. J Biomed Ma ter Res 2001;54(4):470-9. 131. Wanka G, Hoffmann H, Ulbricht W. Phase Diagrams and Aggregation Behavior of Poly(oxyethylene)-Poly(oxyp ropylene)-Poly(oxyethylene) Triblock Copolymers in Aqueous Solutions. Ma cromolecules 1994;27(15):4145 4159. 132. Oh SH, Kim JK, Song KS, Noh SM Ghil SH, Yuk SH, Lee JH. Prevention of postsurgical tissue adhesion by ant i-inflammatory drug-loaded pluronic mixtures with sol-gel transition behav ior. J Biomed Mater Res A 2005;72(3):30616. 133. Fu Y-J, Shyu S-S, Su F-H, Yu P-C. Development of biodegradable copoly(-lactic/glycolic acid) mi crospheres for the controll ed release of 5-FU by the spray drying method. Colloids Surf ., B Biotinerfaces 2002;25:269-279.
123 Appendix A: Protocol Preparation Procedure of Coil-type Glucose Sensors A.1. Coiling of Platinum-iridium (Pt-Ir) Wires Cut 0.125 mm Pt-Ir wires into 4-7 cm long. Remove the top Teflon tube (1 cm). Polish the bare wire with a swab in toothpaste. Ultrasonic cleaning the platinum surface in pure water for 5 min. Coil the stripped wire around a 30 G1/2 needle. Ultrasonic cleaning the platinum surf ace in pure water fo r 5 min. again. Carefully pass through a cotton thread t he coils then cut the two ends of the thread, do not let any cotton silks leave out the coils. A.2. Enzyme Coating Glucose oxidase solution preparation: 300 L pure water 12 mg Bovine serum albumin (BSA) 2.5 mg glucose oxidase (GOD) 4 L glutaraldehyde (50% v/v) Dip-coatings (3 times). Let it dry 1 hour at room temperature.
124 Appendix A (Continued) A.3. Epoxy-PU Coating Coating solution preparation: Tetrahydrofunan (THF) 4 mL PU 45 mg Brij30 5mg Epoxy adhesives 50mg Dip-coatings (3 times) & Dry at room temperature for 30-60 min. Coat two-end of coil & Dry at room temperature for 30-60 min. Cure at 80-120C for 60 min. Place in PBS prior to use (at l east 1 day for membrane swelling).
125 Appendix B: Protocol Measu rement of Sensor Function B.1. Preparation of Measurement Testing solution: PBS 5 mM glucose/PBS solution 100 mM glucose/PBS solution Potentiostat options: Select chronoamperometry Set applied potential at 0.7V Cell setup: 8 ml of 5 mM glucose/PBS in a 10 mL glass beaker Connect counter and reference cl amps to the Ag/AgCl electrode Connect working electrode to the glucose sensor B.2. Response Time and Slope Measurement Run the program until the current ( I5mM) reach a stable level. Add 941 L of 100 mM Glucose/PBS into the cell and continue to record the current change until the second current ( I15mM) level stable. Response time may be expressed as T95%(sec.) [Fig. B.1]. Sensitivity (S) can be roughly calculated by: S (nA/mM) = ( I15mM I 5mM) / (C15mM C5mM) = I / 10
126 Appendix B (Continued) Time (sec.)Current (nA)T95%I5mMI15mM95% I Max. I Time (sec.)Current (nA)T95%I5mMI15mM95% I Max. I Figure B.1. Amperometric Response Curve. B.3. Preparation of Calibration Plot Run the program until the backgr ound current reach a stable level. Step-add x L of 100mL Glucose/PBS into the cell every 500 sec. Obtained a group of co rresponding currents ( I 1 I 7). Draw the current-con centration dependence. Response sensitivity can be obtained by calculating the slope of the current ( I ) vs g lucose concentration (C) linearity relationship.
127 Appendix B (Continued) Table B.1. Changes of Glucose Concentration in the Cell. X (mL) of 100 mM glucose/PBS solution Glucose concentration (mM) Current (nA) 0.163 2 I1 0.258 5 I2 0.467 10 I3 0.523 15 I4 0.589 20 I5 0.667 25 I6 0.762 30 I7
128 Appendix C: Protocol Implantation of Glucose Sensors in the Rat and Measurement of Sensor Function In vivo C.1. Surgery Materials Male Sprague Dawley Out bred Rats (375g 399 grams). Isofluorane anesthesia machine. Surgical clippers and water circulating heating board. Sterile bench pads, drapes, surgery packs with scalpels, and probes. 3-0 Proline sutures. Sterile and non-sterile gloves. Lab note book, pen and sharpie. Sterile and un-sterile gauze. Puralube ointment for eyes. 50cc of D50 glucose solution. Sterile towels (to keep animal warm). Chlorhexidine / Betadine solutions. Isopropyl alcohol. 50cc of D50 glucose solution. C.2. Glucose Monitoring and Testing Apparatus Apollo 4000 Free Radical Analyzer. FreeStyle Blood Glucose Monitoring System.
129 Appendix C (Continued) C.3. Sterilization and Pre-ca libration of Glucose Sensors All sensors will be sterilized by 70% ethanol. Each sensor will be incubated in a sterile 5mM glucose solution three days before implantation. On the day before implantation, each sensor will be pre-calibrated with 5 mM and 15mM glucose solutions which are st erilized by a sterile syringe filter. After obtaining the sensitivity value of each individual sensor, the sensor will be stored in sterile distilled water. C.4. Protocol for Animal Surgery All surgical instruments and other item s to be sterilized will be autoclaved at 260C for 25 min. doubled wrapped or in the sterilization pouch. Surgery will be conducted on a clean surface wiped with disinfectant before and after use. A Continuous Flow Gas Anesthesia Syst em (flow meter, vaporizer, tubing and connectors) will be used to deliver Isoflour ane to the rats. The animals will first be placed in an induction chamber for i nduction of anesthesia and then the gas will be delivered through a rat mask when surgery is performed. Position the first rat on the wa ter circulating heating board fo r rodents using tape to insure positioning of the body, head and rat mask.
130 Appendix C (Continued) Each of the rats that are under anest hesia will have their ey es lubricated with Puralube ointment. C.5. Implantation of Sens ors (Long Wire Sensors) An area on the dorsal aspect of each rat will be shaved at the cervical region to the lumbar region. The skin will be surgically prepped using 3 scrubs of 2% Chlorhexidine and painted with Betadine and left to dry. A sterile fenestrated drape wil l be placed on the rat. Each rat will have 2 sensors implanted. Using the scissors a 1.5 cm incision is made at the dorsal midline 3cm below the inter-scapular area. Lateral inci sions are made 1cm below the interscapular area 1cm lateral to the dorsal midline on either side. The 14-gauge. I.V. catheter was insert ed subcutaneously toward the incision from the 4 5 cm lower back region. Withdraw the 14-gauge needle leaving the cat heter in place. The sensor can then be carefully adv anced, using thumb and forceps, through the catheter without touching t he distal end of the sensor. The sensor was secured to the skin by passing a 3-0 Prolene suture through the small gap of the wound clip on the sensor wires. The incision was sutur ed using 3-0 Prolene.
131 Appendix C (Continued) The cannula was then retracted, leaving the sensor in the subcutaneous tissue. The sensor was secured to the skin by passing a 3-0 Prolene suture through the small gap of the wound clip on the sensor wires. The incision was sutur ed using 3-0 Prolene. The cannula was then retracted, leaving the sensor in the subcutaneous tissue (For short wire sensors the sensors were direct ly implanted through the incision without using a cannula). C.6. Sensors Testing Sensor testing will be performed with two anesthetized rats at a time. A total of 8 sensors will be tested per day. The implanted sensors wires will be atta ched to the Apollo 4000 potentiostat and a 0.7V vs Ag/AgCl potential will be applied to four sensors. At the same time, four response current curves will be continuously recorded on digital display. After the one hour run-in period is complete a re latively stable signal ( I1) from the sensors will be recorded. Using the Freestyle glucometer the low blood glucose level (C1) will established using 1/3 micro liter of blood from the rat tail.
132 Appendix C (Continued) After a stable signal is obtained from t he sensors, 2.0 g/kg rat body weight of 50% glucose/water solution will be admin istered intraperitoneally using a 27 gauge IV needle. An increase in the blood glucose of t he rat will correspond with a rise in the slope of the current-time curve of each sensor. Previous in vivo studies have demonstrated that plasma glucos e will increase to a plateau ( I2). This time interval is long enough to establis h equilibrium betw een plasma and subcutaneous glucose concentrations. More blood tests will be made every 5-10 minut es after injection of glucose until the high glucose level (C2) is stable. A blood-calibrated sensitivity (S) can be calculated: S (nA / mM) = I2 I1 / C2 C1 The same test will be performed on days 7, 14, 21, and 28 to establish the sensitivity of the sensors over time. C.7. Animal Recovery Remove the animal from the anesthesia device. Gently place the animal back into its cage. Place half the cage onto a circulating heat ing broad or place an electric heating pad in the cage to cover half the area of the cage floor.
133 Appendix C (Continued) Once fully recovered from anesthesia, replace the water, food, but no toys or tunnels in the cage with the animal. Daily observation of the rats with not es in the lab book should be done.
ABOUT THE AUTHOR Young Min Ju received a Bachelor of Science in Macromolecular Science from Hannam University (South Korea) in 1996. He continued to study in the Biomedical Polymers at t he graduate school of his alma mater. While in graduate school years, he was assigned as a re search and teaching assistant under the supervision of Dr. Jin Ho Lee. Afte r earning a Master of Science in Macromolecules in 1999, He worked as a research scientist at the Biomaterials Research Center in the Ko rea Institute of Science & Technology (KIST) until summer 2003. He entered the Ph.D. in Biomedical Engineering Program at the University of South Florida in fall 2003 and worked as research assistant in the Biosensor and Biomaterials Lab. under the supervision of Dr. Francis Moussy.
xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam Ka
controlfield tag 001 002007396
007 cr mnu|||uuuuu
008 090619s2008 flu s 000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0002354
Ju, Young Min.
A novel bio-stable 3D porous collagen scaffold for implantable biosensor
h [electronic resource] /
by Young Min Ju.
[Tampa, Fla] :
b University of South Florida,
Title from PDF of title page.
Document formatted into pages; contains 133 pages.
Dissertation (Ph.D.)--University of South Florida, 2008.
Includes bibliographical references.
Text (Electronic dissertation) in PDF format.
ABSTRACT: Diabetes is a chronic metabolic disorder whereby the body loses its ability to maintain normal glucose levels. Despite of development of implantable glucose sensors in long periods, none of the biosensors are capable of continuously monitoring glucose levels during long-term implantation reliably. Progressive loss of sensor function occurs due in part to biofouling and to the consequences of a foreign body response such as inflammation, fibrosis, and loss of vasculature. In order to improve the function and lifetime of implantable glucose sensors, a new 3D porous and bio-stable collagen scaffold has been developed to improve the biocompatibility of implantable glucose sensors. The novel collagen scaffold was crosslinked using nordihydroguaiaretic acid (NDGA) to enhance biostability. NDGA-treated collagen scaffolds were stable without any physical deformation in the subcutaneous tissue of rats for 4 weeks.The scaffold application does not impair the function of our sensor. The effect of the scaffolds on sensor function and biocompatibility was examined during long-term in vitro and in vivo experiments and compared with control bare sensors. The sensitivity of the short sensors was greater than the sensitivity of long sensors presumably due to less micro-motions in the sub-cutis of the rats. The NDGA-crosslinked scaffolds induced much less inflammation and retained their physical structure in contrast to the glutaraldehyde (GA)-crosslinked scaffolds. We also have developed a new dexamethasone (Dex, anti-inflammatory drug)-loaded poly(lactic-co-glycolic acid) (PLGA) microspheres/porous collagen scaffold composite for implantable glucose sensors. The composite system showed a much slower and sustained drug release than the standard microspheres. The composite system was also shown to not significantly affect the function of the sensors.The sensitivity of the sensors with the composite system in vivo remained higher than for sensors without the composites (no scaffold, scaffold without microspheres). Histology showed that the inflammatory response to the Dex-loaded composite was much lower than for the control scaffold. The Dex-loaded composite system might be useful to reduce inflammation to glucose sensors and therefore extend their function and lifetime.
Mode of access: World Wide Web.
System requirements: World Wide Web browser and PDF reader.
Advisor: Francis Moussy, Ph.D.
Implantable glucose sensor
x Chemical & Biomedical Engineering
t USF Electronic Theses and Dissertations.