UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Bisphosphonate-containing coatings for bone implants Duan, Ke 2007

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_2007-317631.pdf [ 19.37MB ]
Metadata
JSON: 831-1.0078533.json
JSON-LD: 831-1.0078533-ld.json
RDF/XML (Pretty): 831-1.0078533-rdf.xml
RDF/JSON: 831-1.0078533-rdf.json
Turtle: 831-1.0078533-turtle.txt
N-Triples: 831-1.0078533-rdf-ntriples.txt
Original Record: 831-1.0078533-source.json
Full Text
831-1.0078533-fulltext.txt
Citation
831-1.0078533.ris

Full Text

BISPHOSPHONATE-CONTAINING COATINGS FOR BONE IMPLANTS by K e Duan A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENT FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE STUDIES (Materials Engineering) The University of British Columbia August 2007 © Ke Duan, 2007 ABSTRACT Bone implants are extensively used to replace joints affected by skeletal problems. Challenges remain to further improve the clinical outcomes of implants. A fast and strong implant fixation would improve the patient's quality of life and reduce implant failure risk. The service life of implants needs to be extended, particularly for the younger patients. A logical approach to these challenges is to control the peri-implant bone formation and remodeling. Bisphosphonate drugs, potent osteoclast inhibitors, are the appropriate choice for this purpose. This thesis studied bisphosphonate-containing coatings on bone implants for local drug delivery. A reproducible electrolytic deposition (ELD) process was developed to prepare calcium phosphate (CaP) coatings on Ti and Ta as bisphosphonate carrier. The ELD parameters were experimentally determined. Microporous coatings containing octacalcium-phosphate were obtained; the pore sizes were 0.5-1 um. The ELD current showed a rapid Cottrell-type decay followed by a prolonged nearly-constant stage, corresponding to proton reduction and molecular water electrolysis, respectively. Alendronate was chemically adsorbed on the CaP coating, and the in vitro release from the coated porous Ta implant was slow, with -10% released after 7 days. The ELD technique was extended to process solid bisphosphonate coatings. Uniform coatings of calcium-etidronate and calcium-alendronate were deposited on Ti and porous Ta. The ELD process did not alter the molecular structures of the bisphosphonates. The solubility of the coatings in a "physiological" buffer solution was 6 x l O " 5 M for Ca-i i etidronate and 2.5 x10" 4 M for calcium alendronate. In vitro release of alendronate from the calcium-alendronate-coated porous Ta was completed within 3 days, and the alendronate concentration was below the solubility limit. To evaluate the in vivo performance, porous Ta implants coated with the bisphosphonate-containing coatings were implanted into rabbit tibial diaphyses together with control implants. Four weeks after implantation, the CaP-coated implants with chemically adsorbed alendronate showed significantly higher total new bone area and push-out strength. The implants with calcium alendronate solid coating showed similar implant fixation and new bone area to the native Ta implants. Eight weeks after implantation, the differences in push-out strength and total new bone area were not significant among different implants. in T A B L E OF CONTENTS A B S T R A C T ii T A B L E OF CONTENTS iv LIST OF T A B L E S ix LIST OF FIGURES x LIST OF ABBREVIATIONS xvii A C K N O W L E D G E M E N T S xviii C H A P T E R 1 INTRODUCTION 1 C H A P T E R 2 L I T E R A T U R E R E V I E W 3 2.1 B A C K G R O U N D OF JOINT R E P L A C E M E N T 3 2.1.1 Skeletal Diseases 3 2 .1 .2 Joint Replacement 4 2.1 .3 Methods of Fixation 6 2.2 C U R R E N T C H A L L E N G E S 8 2.2.1 Implant Fixation 8 2 .2 .2 Implant Loosening 10 2.3 IMPLANT SURFACE MODIFICATIONS TO IMPROVE FIXATION 13 2.3.1 Plasma Sprayed Calcium Phosphate Coatings 14 2 .3 .2 Biomimetic Coatings 16 2.3 .3 Electrolytically Deposited Coatings 19 2.4 D E V E L O P M E N T OF BIOMATERIALS TO D E L A Y IMPLANT LOOSENING 2 3 iv 2.5 R E M O D E L I N G : A K E Y A S P E C T IN F IXAT ION A N D LOOSEN ING 2 4 2.6 B l S P H O S P H O N A T E S 2 5 2.6.1 Structures and Properties 2 5 2.6.2 Application of Bisphosphonates in Bone Implants 2 9 CHAPTER 3 SCOPE AND OBJECTIVES 33 3.1 SCOPE 3 3 3.2 OBJECTIVES 3 4 CHAPTER 4 ELECTROLYTIC DEPOSITION OF CAP COATINGS FOR LOCAL DELIVERY OF BISPHOSPHONATE 35 4.1 INTRODUCTION 3 5 4.2 M A T E R I A L S A N D M E T H O D S 3 5 4.2.1 Substrates 3 5 4.2 .2 Determination of CaP Precipitation Boundary 3 6 4.2.3 Determination of ELD Potential 3 7 4.2 .4 Electrolytic Deposition of CaP Coatings 3 7 4.2.5 Chemical Adsorption of Alendronate 4 0 4.2 .6 Characterization Methods 4 0 4.3 RESULTS 4 4 4.3.1 Precipitation Boundary and Coating Solutions 4 4 4.3 .2 ' Voltammograms and ELD potential 4 6 4.3.3 Coating Deposition — Effect of Solutions 4 8 4.3.4 Coating Deposition —Effect of Deposition Time 58 4.3.5 Coating Deposition —Effect of Oxygen and Nitrate 62 4.3.6 Chemical Adsorption of Alendronate 63 4.3.7 ELDonTa 66 4.3.8 In vitro Release of Alendronate from Porous Ta 68 4.4 DISCUSSION 69 4.4.1 ELD Mechanisms 69 4.4.2 Coating Phase and Microstructure 71 4.4.3 Alendronate Adsorption and Release 74 4.5 CONCLUSIONS 76 C H A P T E R 5 E L E C T R O L Y T I C D E P O S I T I O N O F C A L C I U M B I S P H O S P H O N A T E C O A T I N G S 77 5.1 INTRODUCTION 77 5.2 MA T E R I A L S A N D METHODS 77 5.2.1 Bisphosphonates 77 5.2.2 Determination of Ca43isphosphonates Precipitation Boundaries 78 5.2.3 Electrolytic Deposition of Ca-Bisphosphonate Coatings 78 5.2.4 Preparation of Reference Precipitates 79 5.2.5 Materials Characterizations 80 5.3 RESULTS 84 5.3.1 Precipitation Boundary and ELD Solution 84 vi 5.3.2 Reference Precipitate 86 5.3.3 Coating Morphology and Composition 86 5.3.4 Coating Chemical Structure 9 0 5.3.5 In vitro Solubility 9 4 5.3.6 In vitro Release 95 5.4 DISCUSSION 9 6 5.5 CONCLUSIONS 9 9 C H A P T E R 6 IN VIVO STUDY OF BISPHOSPHONATE-CONTAINING COATINGS 1 0 0 6.1 INTRODUCTION 100 6.2 MA T E R I A L S A N D METHODS 101 6.2.1 Implants 1 0 1 6.2.2 Animals 1 ° 2 6.2.3 Implantation Procedure 103 6.2.4 Push-out Test 105 6.2.5 Hi stomorphometry 108 6.2.6 Statistical Analyses 1 1 ° 6.3 RESULTS U ° 6.3.1 Push-out Force and Strength H O 6.3.2 Hi stomorphometry 1 1 6 6.4 DISCUSSION I 2 7 V l l 6.4.1 Main Findings 127 6.4.2 New Bone Formation and Implant Fixation 128 6 . 5 CONCLUSIONS 134 C H A P T E R 7 C O N C L U S I O N S A N D R E C O M M E N D A T I O N S 136 7.1 CONCLUSIONS I 3 6 7.2 RECOMMENDATIONS 138 R E F E R E N C E S 142 L I S T O F P U B L I C A T I O N S • 159 A P P E N D I C E S 160 APPENDIX A 160 APPENDIX B 166 APPENDIX C 169 V l l l L I S T O F T A B L E S Table 2.1 Structure and potency of some commercially available bisphosphonates 28 Table 4.1 Conditions of ELD solutions (concentrations in mM) 45 Table 4.2 Summarized band frequencies and relative intensities of infrared spectra of pure CaP substances and the coatings deposited in this study. Al l numbers in cm"1. 56 Table 4.3 Ca/P atomic ratio of coatings by ELD on Ti (-1.15 vs SCE) in coating solution 3 (Ca 2 + 5.25 mM, phosphate 10.5 mM, pH 5.3) for different times (mean, n=3). 59 Table 5.1 Ca/P atomic ratio of Ca-bisphosphonate coatings prepared on Ti by ELD (-1.15 V vs SCE, 3.5 mM bisphosphonate, 7 mM Ca 2 + , pH 4.6 for etidronate and 4.8 for alendronate ) for different times (mean, n=3) 87 Table 6.1 Number of rabbits used and porous Ta implants tested 102 Table 6.2 Summary of push-out and histomorphometric results 113 Table 6.3 List of p values of statistical analyses (ANOVA followed by PLSD test) for push-out force and strength of porous Ta implants after 4 weeks of implantation. 115 Table 6.4 List of p values of statistical analyses (PLSD test after ANOVA) for ingrowth percentage and ingrowth bone area of porous Ta implants 121 Table 6.5 List of p values of statistical analyses (PLSD test after ANOVA) for callus area of porous Ta implants after implantation for 4 weeks 122 Table 6.6 List of p values of statistical analyses (PLSD test after ANOVA) for total new bone area of porous Ta implants after implantation for 4 weeks 123 IX LIST OF FIGURES Figure 2.1 Radiographic images of a hip (left) and a knee (right) joint replaced with artificial joint implants. (Adapted from Wikipedia with permission) 5 Figure 2.2 Morphology of (a) sintered Ti beads, (b) sintered Ti fibers and (c) porous Ta ((a) and (b): reprinted from Ref [10] with permission from Elsevier) 7 Figure 2.3 Survival curves of artificial hips implanted during 1979-1991 and 1992-2005 in Sweden. (Reprinted from Ref [44], with permission from SNHAR) 12 Figure 2.4 A radiograph showing severe osteolysis around a femoral stem; arrows point to locations of bone loss appearing as radiolucent zones. (Reprinted from Ref [45] with permission from Marcel-Dekker) 13 Figure 2.5 Chemical structure of bisphosphonates; R i and R 2 are side groups (see text).25 Figure 2.6 Microradiograph of (left) a normal rat tibia and (right) a tibia from a rat treated with clodronate. In the treated tibia, due to inhibited resorption, the metaphysis showed denser trabecules and became club-shaped. (Reprinted from Schenk RK et al, Calc Tiss Res, 1973, 11: 196-214, with permission from Springer-Verlag.) 26 Figure 4. 1 Schematics of ELD cell. CE: counter electrode, RE: reference electrode and WE: working electrode 38 Figure 4.2 (a) The ELD cell for coating planar samples; (b) Lab-made fixture for ELD on porous Ta 39 Figure 4.3 Chemical derivatization of the amino group of alendronate by FITC for fluorescent imaging 41 Figure 4.4 The reaction of alendronate with the mino group of fluorescamine forming the fluorescent active conjugate to be detected by fluorescent HPLC 43 Figure 4.5 Precipitation boundary of Ca2+-phosphate solutions with Ca-P ratio fixed at 1:2; error bars indicate standard deviation (n=3); circles with labels 1-4 indicate the conditions of coating solution 1 to 4, which are below the precipitation boundary pH by 1 (referred to Table 4.1) 45 Figure 4.6 Linear sweep voltammograms of Ti. (a) LSV in revised coating solutions 1- 4 (solution 1: K + 42 mM, phosphate 42 mM, pH 4.6; solution 2: K + 21 mM, phosphate 21 mM, pH 5.0; solution 3: K + 10.5 mM, phosphate 10.5 mM, pH 5.3; solution 4: K + 5 mM, phosphate 5 mM, pH 6.1); (b) Comparison of LSV in 0.15 M NaOH and 0.15 M NaCl solutions. For clarity, curves were shifted along the current axis 47 Figure 4.7 (a) Current-time plots of ELD of Ti in solution 1 to 4 ((solution 1: C a 2 + 21 mM, phosphate 42 mM, pH 4.6; solution 2: Ca 2 + 10.5 mM, phosphate 21 mM, pH 5.0; solution 3: Ca 2 + 5.25 mM, phosphate 10.5 mM, pH 5.3; solution 4: Ca 2 + 2.5 mM, phosphate 5 mM, pH 6.1); arrows indicate the onsets of current rise, (b) Current-f1/2 plots; the dotted lines indicate deposition time of 2, 4 and 200 seconds. For clarity, plots were shifted along the current axis 50 Figure 4.8 Scanning electron micrographs of the CaP coating deposited on Ti from different coating solutions at -1.15 V vs. SCE for 1 hour in (a, b): coating solution 1 (Ca 2 + 21 mM, phosphate 42 mM, pH 4.6) and (c, d): coating solution 2 (solution 2: Ca 2 + 10.5 mM, phosphate 21 mM, pH 5.0), (e, f): coating solution 3 (solution 3: Ca 2 + 5.25 mM, phosphate 10.5 mM, pH 5.3) and (g, h): coating solution 4 (Ca 2 + 2.5 mM, phosphate 5 mM, pH 6.1) 52 Figure 4.9 XRD spectra of CaP coatings prepared on Ti by ELD at -1.15 V vs. SCE for 1 hour, from coating solution 1 (Ca 2 + 21 mM, phosphate 42 mM, pH 4.6), coating solution 2 (Ca 2 + 10.5 mM, phosphate 21 mM, pH 5.0) and coating solution 3 (Ca 2 + 5.25 mM, phosphate 10.5 mM, pH 5.3). Arrow points to the characteristic diffraction peak of OCP (010) plane at 4.8° 54 Figure 4.10 FTIR spectra of coatings deposited on Ti by ELD at -1.15 vs. SCE for 1 hour, from coating solution 1 (Ca 2 + 21 mM, phosphate 42 mM, pH 4.6), coating X I solution 2 (Ca 2 + 10.5 mM, phosphate 21 mM, pH 5.0), coating solution 3 (Ca 2 + 5.25 mM, phosphate 10.5 mM, pH 5.3) and solution 4(Ca 2 + 2.5 mM, phosphate 5 mM, pH 5.3). Dotted lines in (a) indicate bands with significant differences in intensities 57 Figure 4.11 SEM micrographs showing surface morphologies of coatings on Ti (ELD: -1.15 V vs. SCE) in solution 3 (Ca 2 + 5.25 mM, phosphate 10.5 mM, pH 5.3) for (a-e): 0.5, 1,2,3 and 8 h. (f): cross-section of an 8 hour coating 60 Figure 4.12 XRD patterns of CaP coatings on Ti by ELD (-1.15 V vs SCE) in solution 3 (Ca 2 + 5.25 mM, phosphate 10.5 mM, pH 5.3) for 1, 2, 3, and 8 hours 61 Figure 4.13 FTIR spectra of CaP coatings on Ti by ELD (-1.15 V vs SCE) in coating solution 3 (Ca 2 + 5.25 mM, phosphate 10.5 mM, pH 5.3) for 0.5, 1, 2, 3 and 8 hours 61 Figure 4.14 SEM micrographs of a CaP coating on Ti (ELD: -1.15 V vs SCE, 3 h) after immersion in PBS/alendronate at 37°C for 7 days 64 Figure 4.15 FTIR spectra of CaP coatings on Ti (ELD: -1.15 V, 3 h) before and after immersion in PBS/alendronate at 37°C for 7 days 64 Figure 4.16 XRD patterns of CaP coated Ti samples (ELD at -1.15 V vs SCE in solution 3) before and after immersion in PBS/alendronate at 37°C for 7 days 65 Figure 4.17 Optical (a) and fluorescent (b) micrograph of a CaP coated Ti (Ti-CaP) together with a Ti coated with CaP and adsorbed with alendronate (Ti-CaP-ALN), both after derivatization with FITC (coating condition: -1.15V in solution 3 for 3 h); the coating with alendronate showed green fluorescence whereas the CaP coating alone appeared dark; 65 Figure 4.18 Linear sweep voltammogram of Ta in revised coating solution 3 (K + 10.5 mM, phosphate 10.5 mM, pH 5.3)), at the scan rate of 20 mV/s. Arrow indicates -1.48 V, the potential chosen for ELD on porous Ta 67 X l l Figure 4.19 SEM micrographs of (a-c): a porous Ta implant coated with CaP by two-electrode mode ELD at 2.5 V in coating solution 3 for 3 h. (b) and (c) are the higher-magnification image of the encircled area of (a) and (b), respectively; (d): a strut of the native porous Ta without the ELD CaP coating 67 Figure 4.20 In vitro release of alendronate from CaP coated porous Ta (coating conditions: 2.5 V in solution 3 for 3 h). Error bars indicate standard deviation (n=3). The broken line indicates the amount of alendronate remained on the CaP coating after 7 days of release, as dissolved by HC1 and assayed by HPLC 68 Figure 4.21 Ratio of HP042"and P0 4 3 " ions in the whole phosphate species with pH. Calculated using pKa of phosphoric acid (pKai: 2.16, pKa2: 7.21, pKa3: 12.32)]. For clarity, H3PO4 and H2PO4" are not shown 72 Figure 5.1 Chemical structures of (a) etidronic acid and (b) alendronate; alendronate is a mono-sodium, zwitterionic salt 78 Figure 5.2 Precipitation pH boundary of (a) etidronate and (b) alendronate in the presence of Ca 2 + , at fixed molar ratio of Ca:bisphosphonate = 2:1; error bars indicate SD (n=3) 85 Figure 5.3 SEM micrographs of (a) Ca-etidronate and (b) Ca-alendronate precipitate (preparation conditions see 5.2.4) 86 Figure 5.4 SEM micrographs of Ca-etidronate coatings on Ti plate for (a, b) 15 minutes, (c, d) 30 minutes and (e, f) 60 minutes (ELD conditions: -1.15 V vs SCE, 3.5 mM etidronate, 7 mM Ca 2 + , pH 4.6) 88 Figure 5.5 SEM micrographs of Ca-alendronate coatings on Ti plate for (a, b) 15 minutes, (c, d) 30 minutes and (e, f) 60 minutes. (ELD conditions: -1.15 V vs SCE, 3.5 mM alendronate, 7 mM Ca 2 + , pH 4.8) 89 Figure 5.6 SEM micrographs showing a uniform Ca-alendronate coating on a porous Ta implant; (b) is the higher magnification view of the marked area in (a). (ELD X l l l conditions: - 1.48 V vs SCE, 3.5 mM alendronate, 7 mM Ca 2 + , pH 4.8, 30 minutes.) 90 Figure 5.7 Ion chromatograms of (a) Ca-etidronate, (b) Ca-alendronate ELD coatings on Ti and their respective standards; Peak 1: water of sample plug, 2: CI", 3: C0 3 2 " . (ELD conditions: -1.15 V vs. SCE, 30 minutes, 3.5 mM bisphosphonate, 7 mM Ca 2 + , pH 4.6 for etidronate and 4.8 for alendronate.)92 Figure 5.8 FTIR spectra of (a) Ca-etidronate and (b) Ca-alendronate ELD coatings on Ti and their respective reference precipitates. (ELD conditions: -1.15 V vs. SCE, 3.5 mM bisphosphonate, 7 mM Ca 2 + , pH 4.6 for etidronate and 4.8 for alendronate.) 93 Figure 5.9 In vitro solubility of Ca-bisphosphonate ELD coatings on Ti: concentration in the buffer solution with time; error bars indicate SD (n=3). (ELD conditions: -1.15 V, 3h, 3.5 mM bisphosphonate, 7 mM Ca 2 + , pH 4.6 for etidronate and 4.8 for alendronate) 94 Figure 5.10 In vitro cumulative release of alendronate from porous Ta implants ELD coated with Ca-alendronate coatings for 30 minutes; error bars indicate SD (n=3). (ELD conditions: -1.48 V, 30 minutes, 3.5 mM bisphosphonate, 7 mM Ca 2 + , pH 4.6 for etidronate and 4.8 for alendronate) 95 Figure 6.1 (a) Photograph showing two porous Ta implants placed in a rabbit tibia, (b-c) Two postmortal radiographs showing the top and side view of two porous Ta implants placed in a rabbit tibia 104 Figure 6.2 Photographs of a bone cortex specimen with a porous Ta implant prepared for the push-out test; (a) endosteal view and (b) side view 106 Figure 6.3 • (a) Schematic and (b) photograph of the push-out test; the specimen is supported on a steel washer and the space between the implant and the P M M A resin is filled with cured PMMA; the implant is push from the endocortical side at 0.5 mm/minute 107 X I V Figure 6.4 Procedure of histomorphometry analyses, (a): In the BSE micrograph, the host bone and new bone showed different greyscales; (b): By image analysis (greyscale thresholding), the host bone cortex was replaced with yellow, the new bone was replaced with blue, and the Ta was replaced with white. The area of each color was then measured 109 Figure 6.5 Representative push-out curves of two porous Ta implants after implantation for 4 and 8 weeks respectively; the 8 week specimen displayed an increase in force after the deformation of porous Ta I l l Figure 6.6 (a) Push-out force and (b) push-out strength of different porous Ta implants after implantation of 4 and 8 weeks; * indicates difference significant over Ta, and ** indicates difference significant over Ta-CaP. Statistically significant differences were found in 4 week implants; p values and number of implants (n) are listed in Table 6.3 (see next page). No statistically significant differences were found by A N O V A for 8 week implants; post hoc analysis was not adopted 114 Figure 6.7 BSE micrographs of different implants after implantation for 4 weeks; (a) Ta, (b)Ta-CaP, (c) Ta-CaP-ALN, (d) Ta-CaALN. The new bone appeared porous, and many existed in isolated island forms, especially at the central pores. .119 Figure 6.8 (a) Bone ingrowth percentage and (b) ingrowth area of different implants after . implantation for 4 and 8 weeks; * indicates difference significant over Ta, and ** indicates difference significant over Ta-CaP. The p values and number of implants (n) are listed in Table 6.4 (see next page). No statistically significant differences were found by A N O V A for the ingrowth area in 8 week implants; post hoc analysis was therefore not adopted 120 Figure 6.9 Callus area of different implants after implantation for 4 and 8 weeks; * indicates difference significant over Ta, and ** indicates difference significant over Ta-CaP. The p values and number of implant (n) are listed in Table 6.5 below. No statistically significant differences were found by A N O V A for 8 week implants; post hoc analysis was therefore not adopted. 122 Figure 6.10 Total new bone area of different implants after implantation for 4 and 8 weeks; * indicates difference significant over Ta, and ** indicates difference significant over Ta-CaP. The p values and number of implants (n) are listed in Table 6.6 below. No statistically significant differences were found for 8 week implants by ANOVA; post hoc analysis therefore was not adopted. . 123 Figure 6.11 BSE micrographs of different implants implanted for 8 weeks; (a) Ta, (b) Ta-CaP-ALN and (c) Ta-CaALN. The new bone appeared to be denser than week 4, especially at the superficial pores 126 Figure 6.12 Correlation between implant push-out force and bone area at (a) week 4 and (b) week 8; results from all implant groups are included 131 Figure B . l Calibration curve of fluorescent HPLC for alendronate, with concentration of 0 to 10 |ig/ml (mean ±SD, n=3, error bars too small to be seen). Mobile phase: 97:3 1 mM EDTA (pH 6.5): methanol, 1 ml/min, 3.9X 150 mm C18 column, excitation 395 nm, emission 480 nm 166 Figure B.2 Calibration curve of ion chromatography for alendronate, with concentration of 2.5 to 30 uM (n=l). Eluent: 28 mM NaOH, 1 ml/minute 167 Figure B.3 Calibration curve of ion chromatography for alendronate, with concentration of 10 to 50 uM (mean ± SD, n=3). Assay procedures following A S T M 6901-99. Absorbance measured at wavelength: 725 nm 168 L I S T O F A B B R E V I A T I O N S AgCl/Ag Silver chloride/silver reference electrode A L N Alendronate CaP Calcium phosphate DCPD Dicalcium phosphate dihydrate EDS Energy dispersive spectroscopy ELD Electrolytic deposition FITC Fluorescein isothiocyanate FTIR Fourier transform infrared spectroscopy HA Hydroxyapatite HEPES 4-(2-hydroxyethyl)-1 -piperazineethanesulfonic acid HPLC High performance liquid chromatography IC Ion chromatography JCPDS Joint committee for powder diffraction studies LSV Linear sweep voltammetry M Molar OCP Octacalcium phosphate PBS Phosphate buffered saline P M M A Poly-(methyl methacrylate) SCE Saturated calomel electrode SD Standard deviation SEM Scanning electron microscope TCP Tricalcium phosphate THR Total hip replacement TKR Total knee replacement UHMWPE Ultrahigh molecular weight polyethylene UV-Vis Ultraviolet-Visible XRD X-ray diffraction A C K N O W L E D G E M E N T S I am grateful to my supervisor, Dr. Rizhi Wang, for his guidance, support and inspirations over the course of my graduate program. I would also like to thank my committee members: Drs. Helen Burt and Tom Troczynski for their help and advising. I would like to thank Dr. Andrew Toms for performing the implantations on the animals, Mr. Youxin Hu for the assistance in preparing the retrieved samples and Ms. Karen Long for HPLC assays of alendronate. Thank you to all the lab-mates: Jestine Ang, Vincent Ebacher, Robin Cao, Yuwei Fan, Youxin Hu, Shanshan Lu, David Shih, Leandro de Macedo Soares Silva and Allen Tang for the great time and talks that we had together. Sincere thanks again to Allen Tang for proof-reading my manuscripts. I am grateful to Walton Killam Trust for offering me the Killam memorial pre-doctoral fellowship and Zimmer Inc. for funding part of the research included in this thesis. Finally, I am indebted to my family: my Mom for her patience and support, and my wife, Yuan Xiang, for her unfailing love and emotional support. I am so indebted to my newborn son, Yutong, for bringing me the joy that nothing else can be compared with. Love you my Angel! X V l l l Chapter 1 Introduction About 1.5 million artificial joints are implanted worldwide each year to relieve the pain and restore the function of human joints affected by problems such as diseases, injuries and congenital deformities [1]. In order to improve the patient's quality of life, there are constant pursuits for faster, stronger implant fixation, and longer implant service life. Fixation of uncemented joints relies on the growth of new bone into the implant surface porosity (i.e. ingrowth) to form a mechanical interlock [2]. The amount of bone interlocking with the implant is an important factor. At the early healing stage, stimulation of more bone ingrowth may accelerate and strengthen the implant fixation [3, 4]. Over a long term, maintenance of more bone interlocking with the implant may delay the loosening and extend the service life of the implant [5]. Bone constantly undergoes balanced resorption and formation, but such a dynamic balance can be disrupted by diseases or drugs [6]. Bisphosphonate drugs are potent inhibitors of bone resorption, and have become the first clinical choice for disorders associated with excessive bone resorption, such as osteoporosis and Paget's disease [6]. Bisphosphonates are also the logical choices for managing the bone mass around orthopaedic implants [7]. By reducing the rate of local bone resorption, it may be possible to produce and maintain a larger volume of peri-implant bone mass. Recently, a few studies showed promising preliminary results using bisphosphonates-coated solid implants in animal models [8, 9, 10, 11, 12]. However, little has been done with porous implants, which are important for the uncemented joint implants. In addition, a technique capable of coating the porous implants evenly and particularly with controlled and uniform microstructure has not been developed. The uniform microstructure would be essential for the reproducible loading and delivery of bisphosphonates. The topic of this thesis is the development of bisphosphonate-containing coatings on porous implants to improve bone-implant fixation. First, a coating technique was developed based on a rational choice of the parameters to prepare uniform and microporous calcium phosphate coatings on titanium and porous tantalum substrates. The calcium phosphate coating was then used as the carrier for bisphosphonates. Second, a new type of coating containing higher dose of bisphosphonates drugs was developed. Finally, bone formation associated with porous implants with or without the two types of bisphosphonate-containing coatings was evaluated using a rabbit cortical bone model. 2 Chapter 2 Literature Review 2.1 Background of Joint Replacement 2.1.1 Skeletal Diseases Skeletal system diseases and traumas are prevalent. It is estimated that about 90% of the population over the age of 40 suffers to some extent from degenerative bone diseases [13]. In Canada, nearly 4 million people have osteoarthritis and 1.4 million have osteoporosis [14, 15]. Each year, there are about 23,000 hip fractures and approximately 70% of them are related to osteoporosis; the number of hip fracture is projected to rise to 88,000 by 2041 [16]. In 1998, musculoskeletal diseases cost Canada 16.4 billion dollars, direct and indirect cost combined [17]. With the aging population, the cost is expected to increase. Hip and knee are the two joints most commonly affected by musculoskeletal problems, mainly by a disease known as osteoarthritis. Osteoarthritis is a disorder due to the breakdown of the joint cartilage, which develops into joint pain, swelling, and stiffness [18]. In severe osteoarthritis, where the cartilage is completely worn out, the bones at the joint rub together and the joint flexibility and mobility are totally lost. The exact cause of osteoarthritis is not fully understood and there is no known cure for it [18]. In Canada, osteoarthritis accounts for 87% of hip and 91% of knee replacement surgeries [19]. 2.1.2 Joint Replacement 2.1.2.1 Hip Replacement In the 1960s, Sir John Charnley established the modern total hip replacement (THR), a revolutionary treatment for hips suffering from skeletal problems such as diseases (e.g. osteoarthritis), injuries, or congenital deformities [20]. In THR, the femoral head and neck are removed and a metallic stem is placed into the femoral medullary canal. A metallic cup with a plastic liner is fixed to the acetabulum. The metallic head on the stem articulates against the liner to realize joint mobility. Fig 2.1 shows the radiographic image of a replaced hip. Most current hip stems and cups are made of commercially pure titanium (Ti), titanium-aluminum-vanadium alloy (Ti6A14V) or cobalt-chromium-molybdenum (CoCrMo) alloy [21]. Ti and Ti alloy have excellent corrosion resistance [22], high strength/density ratio (ultimate tensile strength: 240-300 MPa for pure Ti and 825-860 MPa for Ti6A14V; density: 4.51 g/ml for Ti and 4.43 g/ml for Ti6A14V) and are able to form an intimate interface with bone (i.e. osseointegration) [23, 24]. CoCrMo alloy has higher strength (ultimate tensile strength of wrought CoCrMo: 860-900 MPa) and hardness than Ti alloys [23]; it is prevalently used for joint heads [21]. The liner inside the metallic cup is made of ultrahigh molecular weight polyethylene (UHMWPE) [20, 21]. Currently, THR is a routine procedure. Each year, more than 25,000 hips are replaced in Canada and more than 500,000 worldwide [19]. 4 2.1.2.2 Knee Replacement Total knee replacement surgery was also developed for the knees affected by the skeletal problems. In a typical total knee replacement (TKR) (Fig 2.1), the end of the femur is removed and replaced with a shell prevalently made of CoCrMo alloys (the femoral component) [25]. The end of the tibia is removed and replaced with a piece of UHMWPE plateau backed by a stem made of CrCrMo or Ti alloys (the tibial component). The shell articulates against the plateau to realize the joint function. Each year, more than 33,000 knees are replaced in Canada and more than 600,000 are replaced worldwide [19]. Figure 2.1 Radiographic images of a hip (left) and a knee (right) joint replaced with artificial joint implants. (Reprinted from Wikipedia with permission). 5 2.1.3 Methods of Fixation In the early surgical protocols, the implant was fixed to the bone with poly-(methyl methacrylate) cement, which provided immediate fixation upon curing (setting in several minutes, reaching 90% of the maximum tensile strength within 4 hours, and reaching 100% within 24 hours) [20, 26]. This method is known as cemented fixation. However, some authors reported unacceptably high rates of implant loosening, which is now known to be caused by non-ideal cementing techniques and the limited mechanical properties of the cement [27, 28, 29]. In the 1970s, a new method of fixing implants to bone without using the cement, known as uncemented fixation, was developed [2, 30, 31]. It uses porous coatings or rough textures on the implant surface to achieve mechanical interlock with the host bone. Immediately after implantation, the implant is held in place only by friction forces, which loosens with time with the adjacent trabeculae gradually resorbed; it is critical to facilitate new bone growth into the pores to form a firm interlock with the implant [32]. Nowadays, both cemented and uncemented fixations are used clinically. Cemented fixation may be considered more suitable for older patients and uncemented fixation tends to be increasingly used in younger patients [21]. There are a few of types of porous surfaces for the uncemented fixation. Porous surface beads or fiber (Fig 2.2 a, b) are the commonly used designs [2]. They are manufactured by pressing the metallic (CoCrMo, Ti or Ti alloys) beads or fibers onto the implant surface followed by sintering at elevated temperatures to form metallurgical bonding [2, 33, 31]. In order to achieve an ideal interlock with bone, the pores should be fully interconnected and the pore sizes should be > 100 um [2]. 6 (a) (b) (c) Figure 2.2 Morphology of (a) sintered Ti beads, (b) sintered Ti fibers and (c) porous Ta ((a) and (b): reprinted from Ref [10] with permission from Elsevier). Porous tantalum (Ta) (Fig 2.2 c) is a relatively new biomaterial for porous surface design [34]. It is fabricated by chemical vapor deposition of Ta on an open-cell vitreous carbon foam. It has high (up to 80%), uniform and fully interconnected porosity, and also high coefficient of friction against the bone. These unique properties make it an ideal biomaterial to form strong fixation with bone [34]. Porous Ta is diffusion-welded to the artificial joints as the surface coating [35]. In addition, because of its excellent strength (compressive strength: 60 MPa) and good modulus match (~ 1.5 GPa) with cortical bone [36], it is also used in free-standing forms such as acetabular cups in THR and the tibia component in TKR [37]. 7 2.2 Current Challenges Although total joint replacements are widely performed, there are numerous challenges and needs for further improvement. Among them, two aspects may be of particular importance: improved implant fixation and reduced implant loosening. 2.2.1 Implant Fixation A fast and reliable fixation is crucial for artificial joints. A fast fixation allows the patient to resume normal activity sooner and reduces the health care cost. A fast fixation may also reduce the magnitude of relative movement between the implant and the bone [3, 4]. Immediately after its implantation, an implant is fixed to the adjacent bone only by the friction force. This friction fixation is gradually lost with the resorption of the adjacent necrotic bone during the bone healing. Bearing of the patient's body weight causes the implant-bone relative movement. The excessive movement may lead to the formation of fibrous tissue at the bone-implant interface. A fibrous interface lacks a direct bone-implant contact and thus rigidity and strength; it therefore may in return increase the movement, forming a "vicious cycle" [3, 4]. This vicious circle may proceed and lead to an early implant failure. A reliable implant fixation is particularly valuable when the local bone healing is compromised. Jasty et al implanted cylindrical, porous-coated Ti6A14V implants into femoral metaphyses of dogs and subjected the implants to oscillatory rotation of different magnitudes [38]. They found that implants with 0 or 20 |im linear movement intimately 8 contacted the new ingrown bone, which connected to the surrounding bone. In contrast, implants with 40 or 150 u,m movement showed a mixture of new bone and fibrous tissue at the interface, and the new bone was partly (40 u,m movement) or entirely (150 |im movement) separated from the surrounding bone by fibrous tissues. S0balle et al implanted porous-coated Ti in femoral condyles of dogs and subjected the implant to axial sliding relative to the adjacent bone [39]. The implants with 150 or 500 |im sliding formed dense fibrous tissue without direct bone-implant contact, whereas implants without sliding were in contact with new bone. The push-out strength of implants with sliding was dramatically lower than implants without sliding. A bone-implant interface with micro-movement or fibrous tissues is mechanically weak, especially in resisting shear and tension [40]. Such an implant is therefore prone to dislocation and loosening. A loosened implant causes great pain and must be surgically removed and replaced with a new implant; this procedure is known as revision surgery. By a clinical survey of THRs, Freeman et al found that femoral stems migrating at or beyond 1.2 mm/year during the first two years after the replacement are at significantly higher risk of loosening than stable implants, with a sensitivity of 78% [41]. For TKRs, Ryd et al also found that cumulative implant migration of 2.7 mm one year after the replacement, or 3.3 mm two years after, is strongly correlated to later implant loosening, with a predictive power o f -85% [42]. There are also other clinical challenges for the fixation of implants. Immediately after the placement, an implant may encounter gaps to the surrounding bone, especially in the 9 case of bone deficiencies, revision surgeries, or anatomical mismatches. The presence of gaps makes bone healing and implant fixation more difficult and less predictable. Hofmann et al implanted porous, cylindrical Ti6A14V implants into human femoral condyles for an average of 27 weeks, and found that gaps as small as 50-500 u.m failed to be bridged by new bone and were instead filled with fibrous tissues [43]. 2.2.2 Implant Loosening Loosening is another important issue in joint replacements. Each year in Canada, 12% of artificial hips and 6% of artificial knees implanted are for revision purposes, and 54% and 35% of these revisions are due to loosening1 of the previous implants [19]. The service life of an implant before it fails is a critical indicator of its performance. In a few countries, the service lives of artificial joints are tracked and the Swedish national hip arthroplasty registers (SNHAR) is the most comprehensive database in tracking the outcomes. The following text is based on the latest report of SNHAR [44]. After 14 years of implantation, 88.7% of cemented hips survived (i.e. not revised ) whereas 71.9% of uncemented hips survived (Fig 2.3 a-b, blue lines). For the uncemented hip implants, the risk of revision increases sharply 5 years after implantation. For younger patients, the risk of revision increases substantially, irrespective of the fixation methods, and remained to be significantly higher than the cemented implants. For the hip implants placed in patients younger than 50 years, 78.9% of the cemented implants survived 14 years, 1 There are additional 16% hip and 1 2 % knee revisions classified as instability. 2 In SNHAR reports, revision is the synonym of failure, and survival means not revised. 10 whereas only 63.9% of the uncemented implants survived the same duration (Fig 2.3 c-d). More unfortunately, the revised implants are at higher risks of loosening than the previous ones. Therefore, the risk of failure (mostly due to loosening) is higher for the uncemented implants and is particularly high for the younger patients. The loosened implants usually show remarkable loss of bone stock (i.e. osteolysis) around the them, and the loosening is caused by the lack of bone-implant interlock [45, 46]. Fig 2.4 shows a radiograph of severe peri-implant osteolysis. It is now generally accepted that wear particle-induced responses are the main cause of such peri-implant osteolysis [45, 47]. The articulation of a joint head against the liner liberates billions of wear particles per year, many of which migrate to the bone-implant interface. As many as 109 particles per gram of tissue were recorded in tissues around retrieved total knee or hip implants [48]. Macrophages try to clear the particles by phagocytosis. But if the number of particles exceeds the capacity of macrophages to clear them, macrophages will secrete signal molecules (e.g. cytokines) to recruit osteoclasts, the cells responsible for secreting acid to dissolve bone [45]. As a result, the peri-implant bone is resorbed due to the net-increase in the number and activity of osteoclasts. Therefore, to reduce implant failure rate, peri-implant bone resorption should be inhibited. All Cemented Implants alt diagnoses and all reasons for revision All Uncemented Implants all diagnoses and all reasons for revision Younger than SO years OMTMflted implants, 1992-2005 2 3 4 5 I I fears festepei »st( t I M 11 12 S3 N (c) Younger than 50 years unwmmtcd implants, 1992-2005 (d) Figure 2.3 Survival curves of artificial hips implanted during 1979-1991 and 1992-2005 in Sweden. (Reprinted from Ref [44], with permission from SNHAR) 12 Figure 2.4 A radiograph showing severe osteolysis around a femoral stem; arrows point to locations of bone loss, appearing as the radiolucent zones. (Reprinted from Ref [45] with permission from Marcel-Dekker) 2.3 Implant Surface Modi f icat ions to Improve F ixat ion Because of the crucial impact of implant fixation, it is important to modify the implant surface to accelerate and/or strengthen implant4x>ne integration. Surface modification may take many forms, but given that ~70 wt % of the bone is carbonated apatite (a non-stoichiometric hydroxyapatite, Caio(OH)2(P04)6, HA) [49], coating metallic implants with a layer of calcium phosphate (CaP) is a logical choice. The rationale is that due to their chemical similarities, bone may not recognize CaP as a foreign material, and therefore may heal faster and integrate with the implant more firmly. HA has been known to bond directly to bone without an intervening fibrous tissue, and the chemical bond was suggested to exist at the bone-HA interface [50, 51, 52]. Unfortunately, HA has low strength and toughness; its bend strength is 40-120 MPa and its fracture toughness (Ki c ) is 13 1.0-1.5 MPa.m 1 7 2 [53]. The HA coated implants therefore combine the desirable chemical properties of HA and the mechanical properties of the metallic substrate [3]. 2.3.1 Plasma Sprayed Calcium Phosphate Coatings Currently, CaP coatings are extensively produced in the industry by the plasma-spray technique [54, 55, 56, 57]. In this technique, CaP powder is fed into a high temperature plasma flame (e.g. N2, Ar or H 2) and directed toward the implant. The powders, partially melted in the flame, solidify on the implant surface to build up a coating particle-by-particle. The technique is mainly used to prepare HA coatings, but other CaPs such as a- or (3-tricalcium phosphates (Ca3(P04)2, TCP) may also be co-sprayed with HA to modulate the coating properties like reactivity and resorption rate [58]. The composition of plasma-sprayed CaP coatings commonly deviates from the composition of the powder input. At the plasma flame temperature (up to ~15,000 °C) [57], CaP powders may decompose into calcium oxide, calcium pyrophosphate or TCPs (in the case of HA powders) [55, 56]. Carbonate and hydroxyl groups in HA powders, both normal constituents in bone mineral, may be thermally removed (i.e. decomposition) [55]. Additionally, the fast solidification creates an amorphous CaP phase [59]. Al l of these by-product phases have much higher solubility than HA, therefore small variation in their contents may significantly change the coating dissolution rate and potentially affect in vivo performance [60]. Paschalis et al. compared six commercially available and clinically used HA coatings and found the dissolution rates differed by as much as 5 times, even though X -ray diffraction indicated they were all HA [61]. Post-treatments (e.g. heating) were 14 developed to decrease the amorphous CaP content, but it is still difficult to precisely control the coating composition [62]. Despite these variations, positive results with HA coatings were found clinically. In general, H A coated artificial joints accelerated fixation and improved bone-implant contact even in the presence of gaps [63, 64]. A recent survey of 6,652 cases conducted in Norway showed that for patients younger than 60 years, 2% of HA coated hip stems failed after 10 years, whereas over 25% of the uncoated implants failed during the same period[65]. Unfortunately, plasma-spray has intrinsic technical limitations. It can only coat areas that are in the direct line-of-sight of the plasma flame. It is difficult to uniformly coat implants of complex shapes and it is not feasible to coat internal surfaces such as the porous structures now widely used on joint implants [66]. In addition, the interface between the plasma-sprayed HA coating and metallic implants is primarily mechanical interlock with limited chemical bonding [67]. Due to the fast solidification of the powders and the mismatch in thermal expansion coefficient between the CaP coating and implant substrate, the coating has considerable thermal stress and also frequently has extensive cracks [68]. Plasma-sprayed coatings are comprised of fused grains tens of micrometers in size, a feature very different from bone minerals, and their in vivo resorption is generally slow but unpredictable [69, 70]. Therefore, long-term performance of the coated implants remains a concern. New coating technologies are being sought after to address these limitations. Among them, solution based chemical methods hold unique advantages and great promise. 15 2.3.2 Biomimetic Coatings The biomimetic deposition is a non-line-of-sight coating technique. It is based on the heterogeneous nucleation phenomenon [71]. Thermodynamics demonstrates that when a nucleus forms on a substrate (heterogeneous nucleation), it enjoys a lower activation energy and, importantly, higher nucleation rate than when it forms without any substrates (homogeneous nucleation). Therefore, for a material to crystallize from its supersaturated solution, crystallization can preferentially occur on the substrate to form a coating if the substrate is prepared to have a low contact angle with the nucleus and the solution supersaturation is appropriately controlled at a threshold level [72]. This strategy is termed "biomimetic" because living organisms widely use it to form hard tissues: the organic matrices use anionic surface groups to nucleate biominerals (e.g. calcium phosphate, calcium carbonate) [73]. It has been identified that the electrostatic attraction between surface anions and Ca 2 + is the key step for the nucleation of biominerals [73]. The surface of Ti is a passivated layer of oxide, which is only weakly charged at physiologic pH and therefore is not ideal for nucleating CaP [74]. Campbell et al. introduced anionic group to the Ti surface by silanization followed by sulfonation [75]. The sulfonate-terminated surface induced crystallization of octacalcium phosphate (Ca8(HP04)2(P04)4.5H 20, OCP) and HA coatings from supersaturated solutions without clogging the pores of the porous implants [75, 76 ]. Derivatization into carboxyl or phosphate groups were also found effective in inducing nucleation of CaP [77]. Because the biomimetic CaP coatings were formed at near43ody temperature and physiological pH, their composition, morphology and phases were more similar to bone minerals [78]. The 16 biomimetic coatings were commonly doped with M g 2 + , HP0 4 2", or C0 3 2 ", typically nano-sized (i.e. < 100 nm in at least one dimension) and may contain OCP-HA intergrowth [79]. However, the application of the technique is limited because silanes are rapidly hydrolyzed by water and the moisture content in the processing environment should be controlled for a batch production. Aqueous surface treatments were extensively studied. Kim et al. treated Ti and Ti alloys in 10 M NaOH and heat-treated them at 600 °C [80, 81]. The surface was transformed into a mesh-like sodium titanate with sub-micrometer pores. After soaking for 4 weeks in simulated body fluid, a solution of similar inorganic ions concentrations to the human blood plasma, the surface was covered with a continuous coating of nano-sized, carbonated apatite. Similar results were also obtained with Ta but not stainless steel or CoCr alloy, indicating the influence of different surface hydroxyl groups [80, 82]. The mechanism of CaP nucleation was found to be: N a + - H + exchange increased the local pH and supersaturation, the resulting Ti-OH or Ta-OH acted as the nucleation sites for apatite Many studies adopted the approach but modified the protocols to achieve faster coating processes or less caustic treatments. For example, Ohtsuki et al treated Ti with H2O2 and metal halides (TaCl 5, SnCl 2 etc), and the modified surface induced apatite in SBF in 3 days. Basic Ti-OH groups were suggested to be the nucleation sites for apatite [83]. Wen et al. used an acid-base two-step etching and a more supersaturated solution to grow apatite/octacalcium phosphate coatings at ~ 1 urn/hour [84]. The hydrated titanium oxide layer with micro-nano hierarchical topography was believed to accelerate the nucleation of apatite. 17 Interestingly, the biomimetic process seemed to occur in v i v o as well. When the alkali-heat treated Ti was implanted into the tibiae of rabbits, a calcium and phosphorous-rich layer formed at the Ti-bone interface, which directly bonded to the host bone. In contrast, the untreated Ti did not show such a layer, and was walled off the host bone by a fibrous layer [85]. The alkaline treatments are inherently corrosive, and may damage the integrities of finer structures such as porous surface beads, meshes and metallic foams. Recently, a few groups explored biomimetic coating techniques without any corrosive surface treatments [86, 87, 88, 89]. Since the native orthopedic metals have poor nucleating ability, a higher supersaturation must be used. However, the supersaturation must be controlled below the value where extensive nucleation would occur in the bulk solution. Once it happens, crystal growth will quickly consume Ca 2 + and P0 4 3", and the driving force for surface nucleation disappears. The margin is very small and variable, due to foreign impurities in the solution. Despite this, Zitelli et al. developed a reactor with ion concentrations, temperature and pH strictly controlled [86]. Needle-like apatite uniformly coated CoCr implants with porous surface beads. Barrere et al. purged CO2 into the coating solution to an acidic pH, and then control-released CO2 to slowly raise the solution pH. An amorphous CaP grew on Ti6A14V surface and expanded into a continuous apatite coating [87, 88]. Li employed the same principle, but using H C O 3 " decomposition, to realize the slow pH increase and prepared nano-sized apatite coatings on Ti6A14V [89]. However, the technique requires "extremely tight" control over solution parameters [86], which may result in low process robustness. 18 2.3.3 Electrolytically Deposited Coatings Electrolytic deposition (ELD) is an emerging non-line-of-sight, aqueous and low temperature coating technique [90]. It is based on electrode reaction induced pH jump effect. When a current passes through an electrolytic cell, electrons are injected into the solution through cathodic reactions, and drawn out through anodic reactions. The nature of the reactions depends on the electrode potential and the solution conditions. At the cathode, some common reactions may include: 2H + + 2e —> H 2 | (reduction of proton) 2H 2 0 + 2e —> 20H" + H 2 f (electrolysis of molecular water) The electrode reactions increase the pH of the solution adjacent to the cathode, because of either consumption of acid or generation of base. In a solution containing Ca and phosphate species, the local pH jump drives hydrogen-phosphate anions to deprotonate and thus increases supersaturation of CaP. When pH reaches the critical value, CaP crystallizes onto the cathode to form a coating. Deposition may be carried out under constant current, potential or voltage modes, and the choice of parameters strongly affects the phase and morphology of the coatings obtained. Redepenning carried out ELD on a stainless steel from a saturated solution of calcium dihydrogen phosphate (Ca(H 2P0 4) 2) at pH of 3.5 [91]. Electrolysis at constant currents at 1 to 10 mA/cm 2 for 133 minutes resulted in coatings of dicalcium phosphate dihydrate (Ca(HP0 4).2H 20, DCPD). The coating consisted of radiantly nucleated crystal flakes of lengths from a few hundred micrometers to below 10 micrometers. The dominant reaction was suggested to be the electrolysis of water. Therese et al also conducted an ELD study from a concentrated solution of ammonium monohydrogen phosphate ((NH 4) 2HP0 4) and calcium nitrate (Ca(N03)2) [92]. After passing current at 10-25 mA/cm 2 for 5-60 minutes, DCPD coatings with a similar morphology were obtained. However, based on thermodynamics the electrode reactions were proposed to include reduction of nitrate (NO3") into nitrite (N02"): N0 3 " + 2H 2 0 + 2 e ^ N0 2 " + 20H" A few studies aimed at preparing apatite phase by modifying the solutions. Shirkhanzadeh et al prepared Ca(NC>3) 2—(NH 4) 2HP0 4 solutions with calcium concentration ranging from 20 to 0.61 mM and pH from 4.2 to 6.0 (with a constant Ca/P ratio of 1.67) [93]. The solutions were heated to 85 °C and a constant potential of -1.4 V (vs. saturated calomel electrode) was applied on Ti for 2 hours. FTIR spectra showed increasing OH content in the coatings with lowering Ca-P concentrations, suggesting approaching the HA composition. The mechanism was suggested to be reduction of proton. The coating consisted of randomly packed micrometer-scaled flakes. Aimed at depositing HA coating under a near-physiological pH, Rofiler et al performed ELD on Ti in a solution of 0.033 M calcium chloride (CaCl2), 0.02 M ammonium dihydrogen phosphate (NH 4 H 2 P0 4 ) at the pH of 6.4 [94]. After passing constant currents of 0.5-10 mA/cm 2, electron microscopy and infrared spectroscopy showed 20 that amorphous CaP (ACP) spheres were formed first. With longer ELD time, ACP transformed to apatite crystallites <500 nm long and <60 nm wide. The coating comprised of aggregate of the crystallites, and the morphology was not uniform. Peaker et al compared the ELD parameters on the coverage, morphology and phase of the CaP coatings on stainless steel [95]. Experiments were conducted with the two-electrode constant voltage mode. Under the optimal conditions of 2 volt, the resulting apatite coating displayed uniformly cellular morphology with pore size of ~ 500 nm. However, since the ELD was performed with two-electrode mode (where the electrode potential was not established), the parameters were specific to the setup and it was difficult to adapt the parameters to other ELD configurations. The authors reported-that when the electrode spacing changed from 2 cm to 1 cm, at constant electric field, the coating coverage significantly decreased and the cellular morphology was not observed; when the spacing changed to 3 cm, no coatings could be obtained. In the above studies, no experiments were carried out to confirm the electrode reactions responsible for the coating formations. Yen et al investigated the mechanisms by cathodic polarization scan on Ti [96]. Based on the logI~V curves and the coating phases, the reactions at -0.1 to -0.3 V (vs.Ag/AgCl) were assigned to be the reduction of oxygen and H 2 P 0 4 " ( H2PO4" + H 2 0 + 2e H 2 P 0 3 " (phosphite) + 20H"). The reactions at -0.3 to -1.5 V were assigned to discharge of proton or acidic phosphate anions to generate H2 gas, and the reactions at -1.5 to -3.0 V were assigned to the reduction of molecular water. However, the significant reduction of phosphate (V) to phosphite (III) contradicted the 21 study by Baudler et al [97], who found such reduction could not occur in aqueous solutions. In an ELD study employing current densities even as high as 100 mA/cm 2 no phosphite was detected [98]. Therefore, the mechanisms reported in the study by Yen et al remain to be verified. Despite the incomplete understanding of the process, ELD is generally a robust technique because it does not require strict control over the chemistry of the substrate surface or the solution supersaturation. It is particularly suitable for coating implants with complex shapes due to the unique advantage that once the ceramic film has coated a certain part, the local current density will be automatically reduced due to the high resistance of ceramics. The current density will then concentrate on the bare part [90]. This negative feedback in current density typically resulted in uniform coating coverage. Even though the ELD technique has many advantages, the coating-substrate adhesion strength is sometimes not optimal. This may be partly due to the formation of H2 gas at the substrate-solution interface, which weakens the adherence of the nascent CaP phase to the substrate. Scharnweber et al. patented a cathodic-anodic alternate polarization technique to address the problem [99]. CaP crystals were first formed on Ti surface by a cathodic polarization in an electrolyte containing Ca 2 + and PO4 3". The coated Ti surface was then anodized to form an oxide layer at the surface. The CaP crystals were then incorporated into the oxide layer. Alternating the two steps was claimed to result in a graded calcium phosphate/metal oxide surface coating with improved adhesion. 22 2.4 Development of Biomaterials to Delay Implant Loosening In addition to improving the early bone-implant fixation, a main effort to delay implant loosening is the development of new articulating materials of higher wear resistance to liberate fewer particles. They include crosslinked UHMWPE and metal-on-metal designs. When UHMWPE is exposed to a y-ray or an electron beam, the C-H and C-H bond can be cleaved to form radicals. The radicals may combine with each other and the polymer chains become crosslinked [100]. The crosslinked UHMWPE is harder and more resistant to wear than the untreated material. McKellop treated UHMWPE with y-ray and found the wear rate was decreased by 90% at the 9.5 Mrad and approached zero at beyond 20 Mrad [101]. Muratoglu et al. irradiated UHMWPW with a 10 MeV electron beam, and the resulted material showed no measurable weight loss after 20 million cycles of simulated wearing [102]. However, the crosslinked material was also substantially weaker and more brittle. The protocol of Muratoglu et al. decreased the tensile strength by 45% and elongation-at-failure by 20% [102]. The crosslinked UHMWPE liners may fracture more easily. CoCrMo alloy metal-on-metal design (e.g. CoCrMo wearing against CoCrMo) has a lower wear rate and maintains a high strength [103]. Rieker et al conducted a simulated articulation study and reported a linear wear of 25 (im in the first year, followed by a steady rate of 5 jam/year [104]. However, most of the wear particles are smaller than 90 nm, so this design is estimated to liberate approximately 100 times more particles (by number) 23 than the UHMWPE design, although the volumetric wear is much lower than the UHMWPE design [103, 105]. A concern is that the nano-sized particles may migrate into blood vessels and elevate cobalt and chromium concentrations in the blood. Clinically, 1.5 to 100 times higher concentration of cobalt or chromium were recorded in the blood of patients with hip implants of metal-on-metal designs [106, 107]. Long-term exposure to these heavy metals remains as a major concern. 2.5 Remodeling: A Key Aspect in Fixation and Loosening There is clearly a key aspect in both implant fixation and loosening — the need for more bone volume in direct contact/interlock with the implant. This need relates to the phenomenon of bone remodeling. Bone is a dynamic tissue that undergoes continuous and simultaneous resorption and formation [6]. The process, known as remodeling, allows bone to remove micro-damage and maintain its mechanical integrity. Every year, between 2 and 10% of the human skeletal mass is renewed through bone remodeling. Remodeling is performed through the activities of two cells unique in bone: osteoclasts and osteoblasts [108]. Osteoclasts secret acid to dissolve a pit on the bone. Osteoblasts migrate to the pit and lay down organic matrix, which calcifies into new bone. The activities of osteoclasts and osteoblasts are highly coordinated, and normally balanced. The balance, however, can be disrupted by diseases or drugs [6]. Abnormally high or low osteoclast activities may lead to osteoporosis or osteopetrosis (i.e. stony bone). 24 Therefore, a logical strategy to increase the peri-implant bone volume is to manipulate of the local bone remodeling in favor of bone formation. More specifically, by increasing the bone formation/resorption ratio, the bone-implant fixation may be effectively improved and the implant loosening may be delayed. 2.6 Bisphosphonates The most effective method to influence bone remodeling is treatment with drugs. Although there are drugs that increase bone formation, the most successful drugs in clinics to date are a group of drugs that decrease bone resorption — bisphosphonates [6]. 2.6.1 Structures and Properties Bisphosphonate drugs are a family of compounds analogous to pyrophosphate. Their general chemical structure is shown in Fig 2.5. 0 ; O I R1 C -O I - p = o o O R2 Figure 2.5 Chemical structure of bisphosphonates; R\ and R2 are side groups (see text). Bisphosphonate drugs are potent inhibitors of osteoclastic activities [6]. At the cellular level, they reduce osteoclast recruitment and adhesion, and trigger osteoclast apoptosis. At the molecular level, bisphosphonates inhibit osteoclastic activities by 25 blocking the mevalonate pathway. Because of these unique properties, bisphosphonate drugs have become the first clinical choice in the treatment of a number of bone diseases associated with excessive bone resorption, such as osteoporosis, Paget's disease and osteolytic tumor induced bone loss. Fig 2.6 shows an example of the effect of bisphosphonates in inhibiting bone resorption on the tibiae of growing rats. Figure 2.6 Microradiograph of (left) a normal rat tibia and (right) a tibia from a rat treated with clodronate. In the treated tibia, due to inhibited resorption, the metaphysis showed denser trabeculae and became club-shaped. (Reprinted from Schenk RK et al, Calc Tiss Res, 1973, 11: 196-214, with permission from Springer-Verlag.) Due to similar chemical structures with pyrophosphate, bisphosphonates chelate with calcium. Bisphosphonates therefore have long retention period in bone [6]. Clinical studies showed that a short course of bisphosphonate treatment may inhibit bone resorption for months. To study the effect span of alendronate treatment in reducing bone resorption in postmenopausal women with vertebral osteoporosis, Khan et al treated 21 patients with alendronate for 4 days (30 mg/day intravenous infusion) [109]. The bone resorption rate 26 (assayed by urinary concentration of hydroxyproline, a collagen-specific amino acid) decreased from day 1, reached the nadir by 1 week and remained suppressed for 2 years. The spinal bone mineral density increased by ~5% in the first year and remained significantly higher than the pre-treatment value after 2 years. The authors also determined the half-life of alendronate in bone to be as long as 11 years. The effect span of bisphosphonate treatment was found to correlate with the dosage that the patient received. Eekhoff et al treated 157 patients with Paget's disease1 with two olpadronate treatment protocols: an effective dose (total 40 mg, intravenously given over 5 or 10 days) or a high dose (effective dose followed by oral olpadronate 200 mg/day for 15 days) [110]. The authors found that both protocols effectively alleviated the disease but the duration of the alleviation was dependent on the dosage. After 80 months, the therapeutic effect remained in over 50% of the patients given the high dose but remained in only 30% of the patients given the effective dose. It is widely speculated that bisphosphonates may be chemically bound to calcium and buried in the bone mineral [6, 111]. When osteoclasts start to dissolve that piece of bone, buried bisphosphonate molecules are liberated and terminate the resorption. The effect span thus may be positively related to total amount of bisphosphonates buried in bone. ' Paget's disease is localized increase of bone turnover following a local increase in bone resorption. The biochemical markers are serum alkaline phosphatase (the index of bone formation rate) and urinary hydroxyproline (a product of bone destruction) [6]. 27 There are a number of commercially available bisphosphonate drugs with different side groups [6]. By changing the side groups, the anti-resorptive potency (and toxicity) of bisphosphonates can vary widely. Some common bisphosphonates tested on human are listed in Table 2.1. In this thesis, two bisphosphonates are used: etidronate and alendronate. Etidronate is the first clinically used bisphosphonate and is still used to date for Paget's disease, osteoporosis etc both its potency and toxicity are the lowest. Alendronate is commonly used for a variety of bone diseases, mainly for Paget's disease and osteoporosis; it has relatively high potency but relatively low toxicity [6]. Table 2.1 Structure and potency of some commercially available bisphosphonates.a'b Name Ri R 2 Relative Potency Etidronate -OH - C H 3 1 Clodronate -CI -CI -10 Pamidronate -OH - C H 2 C H 2 N H 2 -100 Alendronate -OH -(CH 2 ) 3 -NH 2 100-1000 Risedronate -OH 1000-10,000 Zoledronate -OH \ — / > 10,000 a The table shows only a general chemical structure of the bisphosphonate compounds. Marketed bisphosphonates drugs have different formulations. b A more comprehensive list is available in Ref [6]. ' Indications vary with countries; the description pertains to Canada and/or the USA. 28 2.6.2 Application of Bisphosphonates in Bone Implants As osteoclast inhibitors, bisphosphonates are the logical choice in enhancing implant fixation and delaying loosening [45, 7]. This has been demonstrated by a few studies. Expecting to inhibit particle-induced peri-implant osteolysis, Shanbhag et al replaced the right hip of 24 dogs and delivered wear particles into the hip joints of 16 dogs, among which 8 also received continuous oral alendronate (5 mg/day until sacrifice) [5]. Twenty-four weeks after surgery, 6 of 7 dogs with wear particles developed peri-implant osteolysis (one dog fractured the femur and was excluded). In contrast, in the 8 dogs treated with both particles and alendronate, only 1 developed osteolysis. To demonstrate the effect of alendronate on new bone formation, Zou et al implanted porous Ta and carbon fiber spinal fusion cages between the lumbar vertebrae of pigs [112]. Three months after surgery, histological studies showed significantly higher bone ingrowth in the pigs treated with oral alendronate (10 mg daily) than the control pigs. However, the systemic bisphosphonate administration has intrinsic disadvantages. The oral absorption of bisphosphonates is only 1-6% under fasting conditions, and is negligible after taking food or beverage [6, 113]. Since the absorption rate is sensitive to these variables, the effective dosages may vary. Bragdon et al compared the effect of oral alendronate (5 mg/day) on bone ingrowth into the porous surface on hip replaced dogs, but failed to observe differences from the control animals [114]. More importantly, the effect of systemically applied bisphosphonates is not site-directed, it is distributed in (and thus affects) the whole skeleton. Inhibition of bone resorption at undesirable skeletal sites can be problematic. Recently, unusually high incidence of the necrosis of jaw bone has been 29 reported in patients (many with cancers) who received dental procedures and bisphosphonates (-95% with zoledronate or pamidronate) [115]. Inhibition of bone remodeling was suggested to be responsible for the necrosis. Although no conclusions were made, the label of zoledronate and pamidronate was recently updated to inform patients of the risk. Local delivery of bisphosphonates is a promising technique that can be used to improve the implant performance but without disrupting the remodeling at undesirable skeletal sites. A few groups explored this technique. To prevent alveolar bone resorption after tooth extractions, Denissen et al explored associating bisphosphonates with porous HA implants [116]. The implants were soaked in the drug solution and implanted in tibial metaphyses of rats for 3 months. The drug treated implants displayed normal bone density, vascularization and bone conduction. The authors did not report any advantages or disadvantages associated with the locally delivered bisphosphonate. Meraw et al studied the effect of alendronate on integration of cylindrical Ti dental implants [117]. Alendronate was loaded on HA-coated implants by solution soaking, but the dose loaded was not reported. Bare Ti implants were coated with 2.8 |lg alendronate, but the details of the coating technique were not disclosed. After implantation in the dog mandibles for 28 days, histomorphometry studies showed that alendronate increased the bone contact with the bare Ti implants by 198%) but reduced the bone contact with HA-coated implants by 43%. In a following study, the group repeated the experiment and 30 reported that the alendronate treated implants had higher new bone area around both types of implants [118]. Yoshinari et al prepared HA coating on Ti rods by ion beam mixing and pamidronate was immobilized (i.e. chemically adsorbed) on the HA coating [8]. After implantation in the dog mandibles for 4 and 12 weeks, the bone-implant contact length was significantly higher for pamidronate treated rods. In a following study, the group immobilized pamidronate on Ca 2 + implanted Ti rods [9]. After implantation in the medullary canal (facing bone marrow) of rats for 1-4 weeks, the pamidronate treated rods showed significantly wider new bone, indicating higher bone formation rate. However, ion beam mixing or implantation is also a line-of-sight technique and is not suitable for coating complex geometries. Tengvall et al used a chemical (therefore non-line-of-sight) method to load bisphosphonates on stainless steel screws [10]. First, the screws were coated with cross-linked fibrinogen, then pamidronate was covalently bonded to the fibrinogen, and finally ibandronate was adsorbed to the surface. After implantation in the tibiae of rats for 2 weeks, the drug treated screws displayed 28% higher pull-out force and 90% higher pull-out failure energy than the screws treated only with fibrinogen. However, in this study the control samples did not include the native stainless steel screws (to date, fibrinogen coated bone implants are not widely adopted clinically). In addition, the chemistry of the adsorption of ibandronate to the surface was not well defined. The role of the covalently bound pamidronate in the adsorption of ibandronate was not fully explained. 31 Recently, Peter et al studied the effect of zoledronate dose on implant fixation. Ti6A14V rods were plasma sprayed with HA and soaked in zoledronate solutions of a range of concentration to load 0, 0.2, 2.1, 8.5 and 16 u,g of drug [11]. The rods were implanted in rat femoral condyles for 3 weeks. The pull-out load increased with drug dose from 0 to 2.1 u,g, then decreased at 8.5 and 16 u.g. The pull-out load was explained by, and correlated to, the denser trabecular networks around the implants with lower doses of drug. However, the number of samples was too small for statistical comparisons. In addition, according to ISO standard 10993-6 (Biological evaluation of medical devices) [119], rat is not considered as an experimental model for skeletal implantation of orthopaedic biomaterials, although it may be used in the first phase of studies. So far, only Tanzer et al investigated the in vivo effect of local delivery of bisphosphonate with porous implants [12]. The outer surface of porous Ta cylinder was first plasma-sprayed with an HA coating, then a solution containing 50 u.g zoledronate was pipetted onto the cylinder and dried. The porous Ta cylinders were implanted into medullary canals of dogs. Twelve weeks after implantation, for the drug treated implants the new bone islands occupied 32.2% of the canal space and 19.8% of the implant pores, but for the control implants new bone occupied only 13.8% of the canal space and 12.5% of the implant pores. However, electron micrographs showed that plasma-spray only coated the exterior surface of the porous implants and the internal surfaces remained uncoated. In addition, the study was not directly conducted in a bony tissue (instead, in the bone marrow), and mechanical fixation of the porous Ta implants was not studied. Chapter 3 Scope and Objectives 3.1 Scope The goal of this thesis is to develop bisphosphonate-containing coatings to influence the peri-implant local bone remodeling, achieve faster/stronger fixation and delay the loosening of artificial joints. The focus is on the uncemented joint implants. CaP coating is clinically proved to improve implant fixation and also has high affinity to bisphosphonates. In a recent study, zoledronate loaded onto porous Ta plasma-sprayed coated with CaP demonstrated higher bone formation in the bone marrow environment [12]. However, the plasma-spray technique is line-of-sight, and lacks control over the coating microstructure and phase content. A new coating technique needs to be developed to apply CaP coatings with reproducible and uniform microstructure on porous implants. In this thesis, a non-line-of-sight ELD and reproducible coating technique is developed to prepare uniform and micro-porous CaP coatings on flat Ti and porous Ta implants. The focus is on rational choice of ELD parameters based on understandings of the ELD mechanisms. The CaP coatings are used as the carrier for surface chemical adsorption of alendronate, an amino-bisphosphonate (2.6.2, Table 2.1). Clinical studies indicate that the effect span of bisphosphonates correlated to the doses applied [110]. Therefore, local delivery of higher dose of bisphosphonate may inhibit bone resorption for longer time, and this may help extend the implant service life by 33 mitigating future osteoclastic osteolysis. However, the surface adsorption technique can only load limited doses, due to its two-dimensional limitation. Therefore, the ELD technique is extended to prepare novel solid bisphosphonate drug coatings to load high dose of bisphosphonate drugs for more prolonged anti-resorption effect. The effect of the two bisphosphonate-containing coatings on the bone formation and fixation is evaluated, using porous Ta implants in an animal model. 3.2 Objectives The specific objectives of this study are: 1. To experimentally establish the major parameters controlling the ELD process, including coating solution concentration, pH, deposition potential or voltage, and deposition time; to study the effect of the parameters on coating phase and morphology. 2. To enhance understandings of the ELD process, and to identify the major reactions responsible for the coating formation at different ELD stages. 3. To chemically adsorb alendronate onto the CaP coatings and to investigate the in vitro drug release. 4. To extend the ELD technique to the processing of calcium bisphosphonate coatings on Ti and porous Ta implants, and to evaluate the in vitro drug release properties. 5. To evaluate the in vivo effect of the two bisphosphonate-containing coatings on new bone formation and fixation of porous Ta implants in the rabbit tibia. 34 Chapter 4 Electrolytic Deposition of CaP Coatings for Local Delivery of Bisphosphonate 4.1 Introduction This chapter focuses on rational establishment of ELD parameters for reproducible preparation of CaP coatings with uniform and microporous microstructure, enhancing understandings of the ELD mechanisms, and studying the chemical adsorption and in vitro release of alendronate on the CaP coating. 4.2 Materials and Methods 4.2.1 Substrates Titanium (Ti) and tantalum (Ta) were used as substrates for coating deposition. Ti plates ( 2 0 X 2 0 X 2 mm) were cut from commercially pure Ti sheets (Johnson Matthey, Edmonton, A B , Canada) and manually ground with 800-grit sandpapers. Ta plates ( 2 0 X 2 0 X 0 . 1 5 mm) were cut from commercially pure Ta foils (Goodfellow, Devon, PA, USA). Porous Ta implants (diameter: 3 . 1 5 mm, length: 5 mm, Hedrocel®) were provided by Zimmer (Allendale, NJ, USA). The porous implants were characterized by a trabecular structure with interconnected pores approximately.500 urn in diameter and a volumetric porosity of about 7 5 % . A l l samples were sequentially cleaned in ultrasonic bath of acetone, ethanol and distilled water, each for 1 0 minutes. Care was taken to prevent acetone and 35 ethanol from directly drying on the sample surface. The cleaned samples were kept in distilled water for the subsequent use. 4.2.2 Determination of CaP Precipitation Boundary In order to choose appropriate solutions for ELD of CaP, the precipitating conditions of CaP were determined by titrations. A series of 50 ml solution containing different concentration of Ca(N0 3) 2 (Aldrich) and N H 2 H 2 P 0 4 (Fisher) were titrated with 0.1 M NaOH solutions until the first sign of turbidity was observed, which indicated the inception of precipitation. A l l titrated solutions had the Ca:P ratio of 1:2. Each solution was titrated three times. The pH at the end-point was quickly measured with a calibrated pH meter (Thermo Orion 410, Beverly, MA). The Ca 2 + and phosphate concentrations at the end-point were calculated using the equation: C = C 0 x — eq. 4.1 ° 50 + V K . NaOH where C is the Ca 2 + or phosphate concentration at the end-point, and C 0 is the original concentration before the titration, and V N 3 O H is the volume of NaOH solution used. 36 4.2.3 Determination of E L D Potential In order to choose electrode potentials for ELD, the potential-current relations (i.e. voltammograms) were first experimentally determined by linear sweep voltammetry (LSV) [Appendix A]. For Ti and Ta plates, the cleaned sample was mounted on a lab-made fixture (Fig 4.1- 4.2) and used as the cathode (i.e. working electrode), together with a platinum (Pt) foil (25x25 mm, Alfa Aesar) as the anode (i.e. counter electrode). A saturated calomel electrode (SCE, Aldrich) was used as the reference electrode. A l l electrodes were partially immersed in solutions and connected to a potentiostat (Gamry PCI4/300, Gamry Instruments, Warminster, PA, USA). The potentiostat was controlled through the equipped PHE 200 physical chemistry software. The potential of the Ti cathode was linearly scanned from -0.3 vs. SCE toward more negative electrode potentials at the scan rate of 20 mV/second, and the current passed was monitored by the potentiostat. LSV of the porous Ta implant was determined in a similar manner using the lab-made fixture shown in Fig. 4.2 b to hold the Ta sample. 4.2.4 Electrolytic Deposition of CaP Coatings Most ELD coating experiments were conducted in the three-electrode electrolysis mode. Samples were mounted on their fixtures and partially immersed in the electrolytic cell filled with 40 ml coating solution. To form the coating, a constant potential was applied by the potentiostat for a given time. After the ELD coating procedure, the sample was removed, rinsed repeatedly with distilled water and air-dried. Two-electrode electrolysis coating mode was used to batch-fabricate CaP coatings on porous Ta samples. The porous 37 Ta was mounted on the fixture (Fig 4.2 b) and positioned in 40 ml coating solution along with the Pt anode, but without reference electrode. A constant voltage was applied by a lab DC power (GPS 1830D, Goodwill, Taiwan) between the electrodes to prepare the coating. Figure 4.1 Schematic diagram of the ELD cell. CE: counter electrode, RE: reference electrode, and WE: working electrode. 38 Figure 4.2 porous Ta. (a) The ELD cell for coating planar samples; (b) Lab-made fixture for ELD 4.2.5 Chemical Adsorption of Alendronate The ELD CaP coatings were then used as carriers for alendronate. Coated metallic samples were soaked in phosphate buffered solution (PBS) containing 1.0 X 10"4 M alendronate (C 4Hi 2NNa07P 2 .3H 20, molecular weight 325.1, ChemPacific, Baltimore, MD, USA) in a 37 °C water bath (Fisher Isotemp 120) for 7 days. After soaking, they were rinsed with copious water (5 times, each 5 minutes) and dried in air. The composition of PBS was 1.06 mM K H 2 P 0 4 , 5.60 mM Na 2 HP0 4 (both Calbiochem) and 0.14 M NaCl (Fisher), and the pH was 7.40. 4.2.6 Characterization Methods 4.2.6.1 Scanning Electron Microscopy Microstructure of the coatings was studied with Hitachi S3000N and 2300 scanning electron microscopes (SEM, Hitachi Scientific, Tokyo, Japan). Samples were sputter-coated with Au-Pd (60:40) alloy before the characterization, using a Denton Vacuum Desk II sputtering coater (Moorestown, NJ, USA) for 30 seconds. For elemental composition analyses, samples without the Au-Pd coating were analyzed with the energy dispersive X -ray spectrometer (EDS) attached to the Hitachi 3000N microscope. 4.2.6.2 X-ray Diffraction The phases of coatings were analyzed with a Rigaku X-ray Multiflex diffractometer (XRD, Cu K a , Rigaku, Tokyo, Japan) working at 40 K V , 20 mA and 1 degree per minute. 40 4.2.6.3 Fourier-transform Infrared Spectroscopy Fourier transform infrared spectroscopy (FTIR) was performed for structural characterizations. CaP coatings on flat samples were scrapped off, pressed with KBr into pellets and tested on a FTIR spectrophotometer (Perkin-Elmer 2000, Wellesley, M A , USA) using a 1 cm"1 resolution mode and 50 scans. 4.2.6.4 Fluorescent Imaging The alendronate adsorbed on the coating was visually examined after chemical derivatization of the amino group [120]. The sample was immersed in 0.2 mg.mr 1 fluorescein isothiocyanate solution (Sigma) in an acetone/water solvent (V:V=1:1) and was kept at 4 °C overnight. The sample was removed and rinsed with acetone, water repeatedly and air-dried. The sample was studied with an optical microscope (Eclipse E600, Nikon, Tokyo, Japan) equipped with a FITC B-2E filter (excitation: 465-495 nm, emission: 515-555 nm) and a high pressure mercury lamp. The derivatization reaction is shown below. S NH-R FITC alendronate fluorescent conjugate Figure 4.3 Chemical derivatization of the amino group of alendronate by FITC for fluorescent imaging. 41 4.2.6.5 In vitro Release of Alendronate Three CaP coated porous Ta implants adsorbed with alendronate were used to evaluate the in vitro drug release properties. Each sample was placed in a capped polyethylene microtube containing 2 ml "physiological buffer" solution and was kept in the water bath at 37 °C. The water bath was not shaken to prevent potential mechanical spallation of the coatings. The buffer solution contains 1.5 mM CaCl 2 (Fisher) and 140 mM NaCl and the pH was buffered at 7.40 by 20 mM of 4-(2-hydroxyethyl)-1 -piperazineethanesulfonic acid (HEPES, Fluka). This buffer solution had similar pH, Ca and NaCl concentrations to human blood plasma [121]. A 1 ml solution was collected at regular intervals for up to 7 days (i.e. 12, 24, 36, 48, 72, 96, 120, 144 and 168 h) and at each time point the microtube was replenished with 1 ml buffer solution to keep the volume constant. After day 7, samples were rinsed with distilled water, dried and immersed in 1.5 ml 1 M HCI (Fisher) to dissolve the remaining coatings. The HCI solution was also kept for analyses. Al l solution samples were contained in capped microtubes, sealed, and kept for analyses. Alendronate released from the CaP coatings was assayed by Ms. Karen Long at the Faculty of Pharmaceutical Sciences, The University of British Columbia, using a high performance liquid chromatography (HPLC) method adapted from Ref [122]. A 100 jil volume of each sample solution was diluted in 0.14 M EDTA (ethylenediamine tetraacetic acid) buffer (pH 10, Fisher) and reacted with 500 | i l of 2 mg/ml fluorescamine in acetonitrile solution (both Fisher) with gentle agitation to form a water-soluble fluorescent 42 conjugate. The acetonitrile was extracted off with 1 ml dichlormethane (Fisher), and the aqueous phase containing the conjugate was withdrawn for analysis by HPLC [Waters 717+ autosampler, Water 600 controller pump, Nova Pak CI8 column (4 u.m, 3.9x150 mm), 470 scanning fluorescence detector, mobile phase 97:3 1 mM EDTA (pH 6.5): methanol, 1 ml/min]. The conjugate was eluted at 1.6-1.9 min and detected by its fluorescence (excitation 395 nm, emission 480 nm). The calibration curve was linear between 0-3.08 x 10"5 M (R2>0.99, see Appendix B). The conjugation reaction is shown below. fluorescamine alendronate fluorescent conjugate Figure 4.4 The reaction of alendronate with the amino group of fluorescamine forming the fluorescent active conjugate to be detected by fluorescent HPLC. 43 4.3 Results 4.3.1 Precipitation Boundary and Coating Solutions The end point of CaP precipitation titration was easy to observe and repeatable. The precipitation pH of CaP increased with decreasing Ca-P concentrations, and the increase was sharper when Ca 2 + < 5 mM (Fig 4.5). An appropriate ELD coating solution should be stable (i.e. without spontaneous precipitation) but also should easily precipitate with a change in pH. Therefore, for a given 2+ Ca and phosphate concentration, the pH should be appropriately below the precipitation boundary value. A number of tests were conducted to evaluate the stability of a range of Ca2+-phosphate solutions at different pH. In general, when the solution pH was below the boundary point by ~0.3, CaP precipitated within hours. This was not unexpected, because the boundary curve reflected the critical supersaturation level, but not the thermodynamic solubility. Crystallization may occur below the critical pH, although at a lower rate. Therefore, the pH of coating solutions was consistently chosen as below the boundary by 1.0. At these pH values, the solutions remained stable with precipitation for days. Four solutions covering the concentration range titrated were chosen for subsequent ELD on Ti and Ta, as listed in Table 1. In addition, 150 mM NaCl was added to the solutions to increase the conductivity, and also to stabilize the ionic strength. 44 Phosphate Concentration (mM) 0 10 20 30 40 50 7 1 | 1 1 > r 1 I | 1 ~ :ipitation pH - o 4 0 0_ 5 1 o 2 Ol 1 I I I I I 1 1 1 1 1 1 0 5 10 15 20 25 C a 2 + Concentration (mM) Figure 4.5 Precipitation boundary of Ca +-phosphate solutions with Ca-P ratio fixed at 1:2; error bars indicate standard deviation (n=3); circles with labels 1-4 indicate the conditions of coating solution 1 to 4, which are below the precipitation boundary pH by 1 (referred to Table 4.1). Table 4.1 Conditions of ELD solutions (concentrations in mM) Concentration Ca(N0 3) 2 NH4H2PO4 NaCl pH solution 1 21 42 150 4.6 solution 2 10.5 21 150 5.0 solution 3 5.25 10.5 150 5.3 solution 4 2.5 5 150 6.1 45 4.3.2 Voltammograms and E L D potential LSV was performed in solution 1 to 4 on Ti to determine the appropriate electrode potential for ELD. However, the voltammograms showed low repeatability, especially for potentials more negative than -1.2 V. After each LSV, a very thin layer of deposit was observed on the Ti surface. The deposit was identified as CaP as it was easily removed by diluted HCI but not by H2O. The deposits may have influenced the current by affecting the mass transport to the Ti surface. Therefore the solutions were revised for LSV by replacing the Ca(NC>3)2 with two-fold concentration of potassium nitrate (KNO3) to prevent the deposit, with NO3" concentration unchanged. The revision should not affect the electrode reactions because potassium (K +) and (Ca2 +) are not reducible under the experimental conditions. LSV in the revised solutions generated repeatable results. The voltammograms (Fig 4.6 a) showed similar features: the currents started from minimal level and significantly increased at approximately -0.5 V, reached plateaus at ~-0.9 V (forming a "wave") and increased substantially approximately at -1.2 V. Comparative LSV (Fig 4.6 b) showed that the wave was present in NaCl solution but absent in the 0.15 M NaOH solution, indicating it was attributed to reduction of proton (H+). The wave disappeared in the NaOH solution because its H + concentration was low (i.e. -6.7 X 10"4 M level). The substantial increase at -1.2 V was then assigned to electrolysis of molecular water. The electrode potential for the subsequent ELD was chosen at -1.15 V, corresponding to the current plateau region, because the starting current would show greater stability against slight variations in the electrode potential. -0.4 -0.8 -1.2 -1.6 Potential (Vvs. SCE) (a) Potential (V vs. SCE) (b) Figure 4.6 Linear sweep voltammograms of Ti. (a) LSV in revised coating solutions 1- 4 (solution 1: K + 42 mM, phosphate 42 mM, pH 4.6; solution 2: K + 21 mM, phosphate 21 mM, pH 5.0; solution 3: K + 10.5 mM, phosphate 10.5 mM, pH 5.3; solution 4: K + 5 mM, phosphate 5 mM, pH 6.1); (b) Comparison of LSV in 0.15 M NaOH and 0.15 M NaCl solutions. For clarity, curves were shifted along the current axis. 47 4.3.3 Coating Deposition — Effect of Solutions Coating solution is an important factor in determining the coating phase, composition and microstructure. The effect of coating solution was studied by ELD in solutions 1 to 4 for 1 hour on Ti plates. ELD Current Figure 4.7 shows the currents passed during the ELD process in different coating solutions. The current-time plots are divided into two stages. The currents decreased rapidly in the first ~300 seconds (1st stage), and kept relatively constant until the end of the ELD process (2nd stage). The currents also showed a slight rise in the second stage (arrows in the Fig 4.7 a). The currents showed good linearity with f 1 / 2 from 4 second to 200 second (Fig 4.7 b), and all linear correlation factors (R2) were greater than 0.97. Morphology Coatings with complete coverage were obtained from all solutions but with different visual appearances. The coatings deposited from solution 1 and 2 were white, the one from solution 3 was grey and the one from solution 4 displayed a rainbow-like color, indicating the decrease in thickness. Big and shiny crystals were seen on the coatings obtained from solution 1 and 2 but not on the coatings obtained from solution 3 and 4. Fig 4.8 shows the SEM morphologies of the coatings prepared from the different solutions. The surface of the coating deposited from solution 1 (Fig 4.8 a) presented many plate-like crystals about 100-200 (im long, 50 |im wide and a few |im thick. These crystals 48 appeared to have nucleated from one point and grown in a radiant pattern. These are the typical morphological features of dicalcium phosphate dihydrate (CaHP0 4 .2H 20, DCPD) [123]. The big crystals were weak as some of them were already broken during sample manipulations in the laboratory. At higher magnification (Fig 4.8 b), a microporous CaP layer was observed underlying the big crystals. The coating deposited from solution 2 (Fig 4.8 c, d) also had the big crystals but the density was significantly lower. A microporous underlying layer was also observed. The coating prepared from solution 3 was free of DCPD (Fig 4.8 e, f) but presented a cellular morphology. The coating consisted of plate-like crystallites 0.5-1 |im long and less than 100 nm thick, which packed into a cellular structure with 0.5-1 urn uniform pores. Remarkably, the coating had no cracks over the whole field of view. The coating deposited from solution 4 (Fig 4.8 g, h) displayed a similar morphology but with 0.2 |im level pore size; it was thinner and the scratch marks (made during grinding) on the Ti substrate were clearly seen. 49 3 < 2 c Z3 O < E •*—' c <D 1_ o solution 4 :' solution 3 • solution 2 solution 1 1000 2000 3000 Time (second) (a) 4000 1/time1'2 (second"1'2) (b) Figure 4.7 (a) Current-time plots of ELD of Ti in solution 1 to 4 ((solution 1: Ca 2 + 21 mM, phosphate 42 mM, pH 4.6; solution 2: Ca 2 + 10.5 mM, phosphate 21 mM, pH 5.0; solution 3: C a 2 + 5.25 mM, phosphate 10.5 mM, pH 5.3; solution 4: Ca 2 + 2.5 mM, phosphate 5 mM, pH 6.1); arrows indicate the onsets of current rise, (b) Current-t"1/2 plots; the dotted lines indicate deposition time of 2, 4 and 200 seconds. For clarity, plots were shifted along the current axis. 50 51 (e) (f) "20 (.im .. _E * • • * - WD 8 Q m 5. 0OkY:;'x2 .-Ok;- 20 urn (g) (h) Figure 4.8 Scanning electron micrographs of the CaP coating deposited on Ti from different coating solutions at -1.15 V vs. SCE for 1 hour in (a, b): coating solution 1 (Ca 21 mM, phosphate 42 mM, pH 4.6) and (c, d): coating solution 2 (solution 2: C a 2 + 10.5 mM, phosphate 21 mM, pH 5.0), (e, f): coating solution 3 (solution 3: Ca 2 + 5.25 mM, phosphate 10.5 mM, pH 5.3) and (g, h): coating solution 4 (Ca 2 + 2.5 mM, phosphate 5 mM, pH 6.1). 52 Phase Structure The coating phases were analyzed by XRD (Fig 4.9). The coatings deposited from solution 1 and 2 showed predominant pattern of DCPD (JCPDS 9-77). The coating from solution 1 showed strong peaks at 11.7°, 21.0°, 23.4° and 29.3°, corresponding to the diffraction of (020), (021), (040) and (041) planes of DCPD. The coating from solution 2 also showed strong peaks at 11.7 and 23.4° but no peaks at 21.0° and 29.3°; the absence of these two peaks was due to the smaller number of crystals on the substrate surface so no crystal was appropriately oriented to meet the diffraction requirement of the respective lattice planes. A weak peak at 26.0° was observed on both patterns and should be attributed to the microporous layer underlying the DCPD crystals. The peak could be indexed to the (002) plane of HA (JCPDS 09-432) or an overlapped diffraction of (2,-1,1) and (002) planes of octacalcium phosphate (CagfHPCMMPO^.S^O, OCP, JCPDS 26-1056). The lattice parameters of OCP and HA were close and their XRD patterns were similar. It was difficult to distinguish the crystal structure of OCP from HA, especially when the crystal size was small or when both were present. The coating prepared from solution 3 displayed a different X R D pattern. It showed a peak at 26.0° but no peaks of DCPD, in agreement with the SEM results. However, the peak at 4.8° (arrow in Fig. 4.8) was the characteristic diffraction of OCP (010) plane, which was the strongest peak in the powder diffraction pattern of OCP. HA had no diffraction peaks below 10.8°. This indicated the coating contained OCP, but the co-existence of HA could not be excluded (due to the strong overlapping). The coating deposited from solution 4 was too thin and did not generate any noticeable diffraction peaks. 53 -4—> c <n 2 9 (degree) Figure 4.9 XRD spectra of CaP coatings prepared on Ti by ELD at -1.15 V vs. SCE for 1 hour, from coating solution 1 (Ca 2 + 21 mM, phosphate 42 mM, pH 4.6), coating solution 2 (Ca 2 + 10.5 mM, phosphate 21 mM, pH 5.0) and coating solution 3 (Ca 2 + 5.25 mM, phosphate 10.5 mM, pH 5.3). Arrow points to the characteristic diffraction peak of OCP (010) plane at 4.8°. Additional structural information was obtained from FTIR and is shown in Fig 4.10. They were compared with the published FTIR spectra of pure CaP substances [124, 125, 126, 127, 128], as listed in Table 4.2. The coating deposited from solution 1 (Fig. 4.10 a, Table 4.2) showed spectra of typical DCPD features. The bands at 1121-1149, 1080-1050, 1005 and 987 cm"1 were assigned to P-O stretching of HP0 4 2 " , and the bands at 578 and 533 cm"1 to P-0 bending of HPO42". The bands at 874 and -1220 cm"1 were assigned to P-OH stretching and O-H 54 bending of H P O 4 2 " , respectively. The band at 795 cm"1 was assigned to OH stretching of water, and the band at 670 cm"1 to the vibration of water. The coating deposited from solution 2 showed similar features with the solution 1, but with a few noticeable differences. The coating from solution 2 exhibited lower intensities at 1218, 876 and 795 cm"1, but developed two new bands at 603 and 562 cm"1. The decrease at 1218 and 876 indicated lower P-O-H content. The two new bands were absent in DCPD but instead could be found in the spectra of HA and OCP (Table 4.2), indicating their increase in content. These changes were consistent with the SEM and XRD characterizations. The coating prepared from solution 3 showed the strongest absorption band at 1039 cm"1, a shoulder at 1077 cm"1 and a weak band at 961 cm"1, all assigned to the P-0 stretching of PO4 3 ". The strong bands at 601, 561 and a shoulder at 576 cm"1 were assigned to the bending of P0 4 3". A strong band at 1106 cm"1 was assigned to the P-0 stretching of H P O 4 2 " . Two weak bands at 904 and 871 cm"1 were also assigned to the P-O stretching of H P O 4 2 " , and their presence was characteristic of OCP [129]. Therefore, the FTIR spectrum also suggested the coating to be OCP, in agreement with the XRD result. The coating prepared from solution 4 generally displayed similar band features with the coating from solution 3, suggesting a close structure. However, the bands in the 1000-1200 cm"1 region were more poorly resolved, indicating the crystal structure of the CaP coating to be more disordered or contains more defects. 55 Table 4.2 Summarized band frequencies and relative intensities of infrared spectra of pure CaP substances and the coatings deposited in this study. A l l numbers in cm"1. Pure substances Coatings deposited for 1 h: DCPD OCP HA Solution 1 Solution 2 Solution 3 Solution 4 1280 w 1217s -1220s 1218 m 1190w 1135 s 1121-1149 s,b 1138-1126 s,b 1105 s 1092 s 1075 s 1075 s 1078 sh 1077 sh -1110 s, b 1060 s 1055 s 1080-1050 b, s 1060 s 1035 s 1040 vs 1039 s 1039 s -1035 s,b 1025 s 1005 sh 1006 sh 988 s 987 s 988 s 962 w 962 w 961 w 961 w 910w 904 w 875 s 874 s 876 m 901-878 w 865 w 871 w 790 s 795 s 794 m 660 m 668 m -666 w 630 w 631 m -626 sh -627 sh 599 m 601 m 603 m 601 s 601 s 577 m 575 w 575 m,sh 579 m 578 s 576 sh 559 m 561 m 562 s 561 s 561 s 526 m 525 w 533 m -526 sh Notations: b — broad m — medium s — strong sh — shoulder v — very w — weak 56 I I I I I 1 I I I 1 1 1400 1200 1000 800 600 400 Wavenumber (cm1) (a) po43' i i i i i i i i i 1200 1000 800 600 Wavenumber (cm 1) (b) Figure 4.10 FTIR spectra of coatings deposited on Ti by ELD at -1.15 vs. SCE for 1 hour, from coating solution 1 (Ca 2 + 21 mM, phosphate 42 mM, pH 4.6), coating solution 2 (Ca 2 + 10.5 mM, phosphate 21 mM, pH 5.0), coating solution 3 (Ca 2 + 5.25 mM, phosphate 10.5 mM, pH 5.3) and solution 4(Ca 2 + 2.5 mM, phosphate 5 mM, pH 5.3). Dotted lines in (a) indicate bands with significant differences in intensities. 57 4.3.4 Coating Deposition —Effect of Deposition Time Deposition time is also an important factor in the ELD process. During the ELD process , the local concentration of ions involved in CaP crystallization may change as the overall result of coating growth and ion supply, and this change may affect the coating properties. In order to understand and optimize the influence of time, Ti samples were ELD coated for different times and the coating microstructure, composition and phase structure were studied. Solution 3 was used for the study because of the absence of DCPD and the uniform microporous morphology described in the previous section. The solution 4 was not used because the deposition rate was too low. The morphologies of coatings changed with ELD time (Fig 4.11). The coating deposited for 0.5 hour showed a cellular morphology constructed of interconnected crystallites about 500 nm long and <100 nm thick. The coating deposited for 1 hour showed similar morphologies but the crystallite "walls" of the cells appeared to be slightly thicker and longer. The 2h coating retained the cellular morphology but the pore size increased to about 1.5 (im. Most "cell walls" at this stage appeared to be somewhat curved. After 3 hours of deposition, the crystallite "walls" were more extended and approached the morphology of flakes. The size of crystallite grew to about 2 [tm wide.. After 8 hours, the morphology of the crystals fully extended into flakes. The Ca/P atomic ratio of the coatings continuously increased from 0.5 to 8 hours (Table 4.3). The: coatings presented no cracks and therefore cross-section could not be revealed. To examine the coating thickness, CaP coatings were deposited on Ti foils using the same procedure and then the foils were bent to induce cracks. The cross-sectional morphology of the 8 hour coating (Fig 4.11 f) also 58 showed the increase of pore and crystal sizes from the bottom to the surface and the coating thickness was estimated to be ~ 15 Jim. The thickness of coatings prepared by ELD for 0.5, 1, 2 and 3 hour coating was estimated to be 1, 2.5, 5 and 7 u,m, respectively. The XRD patterns of the coatings are shown in Fig 4.12. The peaks of OCP (4.8°, 9.6°, and 16.3°) increased in relative intensity with ELD time, corresponding to growth of the coating thickness. On FTIR (Fig 4.13), all coatings showed generally similar patterns, but slight variations with deposition time could be noticed. For 0.5 hour coating, the P-0 stretching bands (1000-1200 cm"1 region) and the two P-0 bending bands (561, 601 cm-1) were poorly split, but with longer ELD time the bands became increasingly resolved. This change indicated a gradual decrease of stoichiometric and/or structural imperfection [130]. In addition, the intensity of weak bands at -1290, 1195, 963, 912 and 864 cm"1 developed with longer deposition time. These bands were attributed to P-0 stretching of H P O 4 2 " and the latter two were regarded as the signature of OCP [129]. These evidences suggested the increase of OCP content with deposition time. Table 4.3 Ca/P atomic ratio of coatings by ELD on Ti (-1.15 vs SCE) in coating solution 3 (Ca 2 + 5.25 mM, phosphate 10.5 mM, pH 5.3) for different times (mean ± SD, n=3). E L D Time (h) Ca/P ratio 0.5 1.09 ±0.04 1 1.13 ±0.02 2 1.20 ±0.03 3 1.24 ±0.01 8" • 1.33 ±0.03 59 (e) (f) Figure 4.11 SEM micrographs showing surface morphologies of coatings on Ti (ELD: -1.15 V vs. SCE) in solution 3 (Ca 2 + 5.25 mM, phosphate 10.5 mM, pH 5.3) for (a-e): 0.5, 1, 2, 3 and 8 h. (f): cross-section of an 8-hour coating. 60 5 10 15 20 25 30 35 26 (degree) Figure 4.12 XRD patterns of CaP coatings on Ti by ELD (-1.15 V vs SCE) in solution 3 (Ca 2 + 5.25 mM, phosphate 10.5 mM, pH 5.3) for 1, 2, 3, and 8 hours. I • i , i . 1 . 1— 1400 1200 1000 800 600 Wavenumber (cm 1) Figure 4.13 FTIR spectra of CaP coatings on Ti by ELD (-1.15 V vs SCE) in coating solution 3 (Ca 2 + 5.25 mM, phosphate 10.5 mM, pH 5.3) for 0.5, 1, 2, 3 and 8 hours. 61 4.3.5 Coating Deposition—Effect of Oxygen and Nitrate The electrode reactions responsible for the local pH jump and the formation of CaP (and other ceramics) have been suggested to be the reduction of 0 2 and/or NO3", in addition to the reduction of H + and H 2 0, but little evidence were reported [92, 131]. In this study, the reductions of H + and H 2 0 have been identified unambiguously by LSV (Fig 4.6). To fully understand the ELD mechanisms, the roles of 0 2 and/or N03" were studied. To investigate the role of 0 2, a few ELD was performed at more positive than -0.6 V (where oxygen was reduced but the contribution of H + reduction was small) for 1 h, no coating was formed. Therefore the contribution of 0 2 was insignificant. The effect of N03" was studied by comparative experiments on Ti. First, the coating solution 3 was modified by replacing Ca(N0 3) 2 with CaCl 2 to exclude the contribution of N03". ELD conducted in the modified solution for 1 hour resulted in uniform coatings of good coverage, indicating that the presence of N0 3" was not essential. Second, to determine whether N03" was actually cathodically reduced, LSV was performed in 0.1 M NaOH solutions with and without 0.1 M KN0 3 . The NaOH solution was used to remove the background of H + reduction at more positive than -1.2 V (vs. SCE). No reduction current could be identified at the potentiaf of more negative than -1.2 V. At more negative than -1.2 V, comparisons could not be made because of the background of exponentially increasing electrolysis of molecular water. However, it can be concluded that at the ELD potential used in this study (-1.15 V on Ti), no evidence of N03" reduction was found. 62 4.3.6 Chemical Adsorption of Alendronate The uniform and microporous coatings were then used as the carrier for alendronate. Coatings deposited for 3 hours on Ti were soaked in the PBS/alendronate solution for 7 days to adsorb the drug onto the surface of the CaP crystals. The coatings before and after the adsorption were analyzed and compared. The morphology of the coating did not show significant change after the adsorption (Fig 4.14). Only minor signs of surface erosion were observed. After the adsorption, the coating exhibited essentially identical FTIR (Fig 4.15); the intensities of HP0 4 2" bands (arrows in Fig 4.16) did not change and the characteristic bands of OCP (912, 860 cm"1) were kept. On XRD, the diffraction at 4.8° (OCP 010 plane) was observed before and after the adsorption; no significant change in peak intensities or new peaks were observed. Therefore, these characterization techniques did not detect any remarkable changes in the coating properties by the adsorption procedure. However, the presence of alendronate on the coating was observed by fluorescent microscopy after conjugated with FITC (Fig 4.17). The coating after the chemical adsorption procedure showed strong green fluorescence, indicating the presence of amino group of alendronate. In contrast, the native CaP coating and the uncoated Ti parts were dark. The green fluorescence was observed over the entire sample and was found to be qualitatively uniform, indicating alendronate was uniformly adsorbed on the coating surface. 63 Figure 4.14 SEM micrographs of a CaP coating on Ti (ELD: -1.15 V vs SCE, 3 h) after immersion in PBS/alendronate at 37°C for 7 days. I , i , i , i , i l 1400 1200 1000 800 600 Wavenumber (cm 1) Figure 4.15 FTIR spectra of CaP coatings on Ti (ELD: -1.15 V, 3 h) before and after immersion in PBS/alendronate at 37°C for 7 days. 6 4 to c CD -4—' c 10 15 20 25 2 9(degree) 3 0 3 5 Figure 4.16 X R D patterns of CaP coated Ti samples (ELD at -1.15 V vs S C E in solution 3) before and after immersion in PBS/alendronate at 37°C for 7 days. Ti-CaP-ALN i Ti-CaP I (a) (b) Figure 4.17 Optical (a) and fluorescent (b) micrograph of a CaP coated T i (Ti-CaP) together with a T i coated with CaP and adsorbed with alendronate (Ti -CaP-ALN) , both after derivatization with FITC (coating condition: -1.15V in solution 3 for 3 h); the coating with alendronate showed green fluorescence whereas the CaP coating alone appeared dark. 65 4.3.7 E L D o n T a To verify the E L D method as a non-line-of-sight technique, and also to prepare implant specimens for subsequent studies, cylindrical porous Ta implants were used as the substrate. To determine the E L D potential for porous Ta, L S V was performed. The voltammogram of Ta in the coating solution 3 (Fig 4.18) displayed different features from T i . The reduction of H + took place at approximately -1.1 V , and the reduction of molecular H2O started at approximately -1.5 V , both more negative than on Ti . In addition, the plateau region was much narrower than Ti . Nevertheless, -1.48 V was chosen as the E L D potential according to the same criteria as in E L D on Ti . E L D in the solution 3 for 3 hours resulted in uniform coatings on porous Ta. During the E L D process, the voltage applied between the Pt anode and the porous Ta was monitored. The voltage was found to be relatively stable at 2.5 V during the E L D process. This voltage was used as the parameter for the following two-electrode E L D process. Two-electrode E L D process was used to batch-fabricate CaP coatings on porous Ta implants with the same fixture. Two-electrode constant voltage E L D was used because of the much wider availability of conventional power sources than potentiostats in real industrial practice. It enables large scale production at low facility costs. Porous Ta samples were E L D coated in coating solution 3 at 2.5 V for 3 hours, and uniform coatings with the similar cellular morphology were successfully formed (Fig 4.19). -0.4 -0.8 -1.2 -1.6 -2.0 Potential ( V v s . S C E ) Figure 4.18 Linear sweep voltammogram of Ta in revised coating solution 3 (K + : 10.5 mM, phosphate 10.5 mM, pH 5.3)), at the scan rate of 20 mV/s. Arrow indicates -1.48 V, the potential chosen for ELD on porous Ta. Figure 4.19 SEM micrographs of (a-c): a porous Ta implant coated with CaP by two-electrode mode ELD at 2.5 V in coating solution 3 for 3 h. (b) and (c) are the higher-magnification image of the encircled area of (a) and (b), respectively; (d): a strut of the native porous Ta without the ELD CaP coating. 67 4.3.8 In vitro Release of Alendronate from Porous Ta Porous Ta implants E L D coated with CaP and adsorbed with alendronate showed a slow and relatively steady drug release (Fig. 4.20). About 1.1 | ig alendronate was released in 7 days, which was 10% of the total amount of alendronate (10.7 ± 3.3 iig) adsorbed on the CaP coatings. After 7 days of release, the samples were examined with an optical microscope; the CaP coating was found to be intact on the Ta surfaces, with only minor signs of local degradation. A 120 100 CD w CO 0) CD CC CD > '•4—» m E zi O 3 4 5 Time (days) Figure 4.20 In vitro release of alendronate from CaP coated porous Ta (coating conditions: 2.5 V in solution 3 for 3 h). Error bars indicate standard deviation (n=3). The broken line indicates the amount of alendronate remained on the CaP coating after 7 days of release, as dissolved by HCI and assayed by H P L C . 68 4.4 Discussion 4.4.1 E L D Mechanisms The ELD of CaP coating was the result of a sequence of electrochemical, acid-base and precipitation reactions. Comparative LSV (Fig. 4.6) in revised coating solutions, NaCl and NaOH solutions indicated the electrode reactions to be: the reduction of (onset at -0.5 V on Ti and -1.1 V on Ta) and molecular H 2 0 (onset at -1.2 V on Ti and -1.5 V on Ta). Although the two reactions are thermodynamically equivalent, they have different reaction paths and activation energies; therefore they were differentiated by LSV. On LSV, the current of H + reduction developed plateaus with potential scans, indicating H + was rapidly consumed (or, "depleted") at the electrode surface and the current gradually became limited by the diffusion of H + to the substrate [132]. The current dramatically increased when the potential reached the onset of H 2 0 reduction, following Tafel's exponential law [133, also see Appendix A]. As the bulk solvent, H 2 0 was not depletable and the current was controlled by the potential. It was not unexpected that Ti and Ta showed different onset potentials for both H + and H 2 0 , since the electrode reaction kinetics was highly dependent on the electrode materials. In this study, the ELD potentials were chosen to be within the plateau region but close to the onset of H 2 0 reduction (Fig 4.6, 4.18). At the chosen potentials, both reactions took place. In the initial stage, H + reduction was the main reaction. But the ELD current decayed over time with depletion of H + , and because the solution was unstirred, the current decay followed Cottrel equation [134, also see Appendix A], thus explaining the linearity 69 between the current and t"1/2 during the first approximately 200 seconds. It might be argued that the formation of CaP may also have contributed to the rapid current drop at the initial stage by increasing the interfacial conductivity; however during the first 200 seconds virtually no coating could be observed by SEM, indicating diffusion controlled H + was the major reason for the current drop. In a later stage, the contribution of H + reduction became sufficiently small and the H 2 0 reduction became the main electrode reaction, corresponding to the relatively constant current stage lasting to the end of the ELD process. We note that at this stage, the ELD current also decreased, but more slowly. Therefore, the ELD mechanisms could be summarized as: 1) consumption of H + by reduction provided the initial supersaturation for CaP deposition, and 2) the coating deposition was sustained by OH" generated from electrolysis of molecular water. In addition to the two main features, the current-time plots (Fig. 4.7) also exhibited slight rises in the currents often after 300 seconds. This may be explained by the reduction of the surface oxide film thickness (or alternatively, partial reduction of titanium dioxide into titanium sub-oxides), and the consequent drop of resistance at the Ti-solution interface. It was also supported by the observation that when the Ti samples were immersed in diluted HF, after removal of the CaP coating with HC1, the part that underwent the ELD process was the first to react with HF to form H 2 bubbles. Montero-Ocampo et al. recently reported a similar, but more dramatic rise of current during ELD of CaP on Ti6A14V alloy, and the extent of the current rise increased with more negative ELD potentials [135]. That phenomenon agrees with our explanation. In addition, other studies also showed that titanium dioxide could be cathodically reduced, although not in aqueous conditions, to suboxides and eventually metallic Ti [136]. The reduction of nitrate (N03~+H20+2e —> N0 2 " + 20H") was suggested to be a main mechanism in the ELD of CaPs and also other ceramic films, based on its relative positive E° (-0.23 V vs SCE) [92, 131]. In this study, however, ELD in nitrate-free solutions demonstrated that the presence of NO3" was not essential. LSV in KNCb/NaOH solution did not provide evidence of nitrate at the working potential in this study. It should also be noted that the relative positive potential of N O 3 7N0 2" couple did not warrant the reduction of NO3" at a certain potential because it was highly dependent on electrode materials. Indeed, Platides et al reported that reduction of nitrate started at -1.6 V (vs. SCE) on SnCu alloy and at -1.4V on A l [137]. At a potential more negative than the present ELD potentials, N0 3 " may be reducible and contribute to the ELD current, but with the exponential increase of H 2 0 reduction current, it remained unproved if it could be the main ELD mechanism. 4.4.2 Coating Phase and Microstructure The coating phase was mainly affected by solution conditions, such as supersaturation and Ca 2 + and phosphate concentrations. The coatings showed three main features. The density of DCPD crystals on the coating surface decreased from solution 1 to 2 and disappeared on coatings deposited from solution 3 and 4. The microporous coatings deposited from solution 3 showed OCP structure and the composition was also close to it (Table 4.3). H A was not identified, but its presence could not be ruled out. 71 Thermodynamic calculations showed that at the respective boundary pH values, coating solutions 1 to 4 were all supersaturated with DCPD, OCP and HA. So the phase of the coatings was not only controlled by the thermodynamics but also by the crystallization rate. The distribution diagram of phosphate species (Fig 4.21) showed that the percentage of HP04 2"in the whole phosphate species increased from pH 5.5 to pH 10. Therefore, the lower C a 2 + and phosphate concentrations, instead of the higher pH, of solution 3 and 4 was the main reason for the gradual disappearance of DCPD. For solution 3 and 4, the Ca 2 + and phosphate concentrations were 1/4 and 1/8 of solution 1, respectively. Under the same pH, the supersaturation of DCPD would be 1/16 and 1/64 of solution 1. In solution 3 and 4, the super saturation levels may become lower than the critical value and the nucleation rate of DCPD became minimal. pH Figure 4.21 Ratio of HP0 4 2"and P0 4 3 " ions in the whole phosphate species with pH. Calculated using pKa of phosphoric acid (pKa,: 2.16, pKa 2: 7.21, pKa 3: 12.32) [138]. For clarity, H3PO4 and H2PO4" are not shown. 72 The formation of OCP (instead of phase pure HA) may be attributed to the interfacial pH under the ELD process, which, however, was difficult to measure. In this study, a qualitative estimation of the interfacial pH was conducted by gradually lowering the pH of the coating solution (from the working pH of 5.3) and performing ELD. It was generally observed that the coverage was significantly lower when the pH was lowered by 1.5, and was negligible when lowered by 2. Although inconclusive, these experiments suggested that the present ELD process only moderately raised the pH of the coating solutions, and less likely to reach extreme pH (i.e. >10). The pH-distribution diagram of phosphate system shows that HPO4 2" was the dominant specie from pH 7.3 up to 10.3 (i.e. 55% and 99%, respectively); in contrast, PO43" has minimal percentage in the pH region (5.3 X 10"4% ~ 0.95%). Because of the predominance of H P 0 4 2 " , which is a constituent of OCP, crystallization of OCP was kinetically favored. In contrast, the paucity of PO43", and also OH", in the pH region limited the crystallization rate of HA. Moreover, H A formed under such conditions would contain relatively high content of HPO4 2" as defects and became difficult to distinguish from OCP by the Ca/P ratio or FTIR. The deposition time affected coating microstructure and composition. With longer ELD, the crystal and pore size increased and the coating morphology changed from cellular to more flake-like. This phenomenon may be explained by the change of supersaturation with time. The crystals on the top (which formed later) were likely formed under lower supersaturation than the ones closer to the substrate (which formed earlier) because the calcium and phosphate ions near the substrate should be gradually consumed by the growth of the coating. The pH, however, may not keep rising because the electrogenerated OH" 73 was also consumed by the coating growth and diffusion to the bulk solution. Indeed, Zhang et al used microsensors to measure the interfacial pH during constant-current E L D of CaPs under different current densities, and found pH rose to the peak values after 5 minutes and slowly decreased with time [139]. A n additional evidence of diminishing supersaturation with E L D is the fact that the E L D coating could not reach infinite thickness; after a certain time, the coating growth stops. The crystal size and coating porosity was determined by the competition of nucleation and crystal growth. Under higher supersaturation, nucleation rate increased more dramatically than crystal growth rate, so the coating formed earlier was constructed of smaller crystals and denser than the surface part. Due to the same reason, the earlier deposits contained more defects and impurities than the later deposits; this explained the lower Ca/P ratio of the coating deposited for shorter times, and also the poor split of the P-O stretching bands on the FTIR spectrum of the 0.5 hour coating (Fig 4.13). 4.4.3 Alendronate Adsorption and Release Alendronate was chemically adsorbed onto the CaP coating from its solution because of its affinity to C a 2 + at the CaP surface. A recent kinetic study showed another bisphosphonate, zoledronate, adsorbed onto H A through phosphonate-phosphate anion exchange (and the release through the reverse process) [140]. Considering the common chemical structure of bisphosphonates family, alendronate should undergo the same process for its adsorption and release from the CaP coating in this study. Because of the stable chelation of alendronate with calcium on the CaP crystal surface, desorption of alendronate 74 requires simultaneous or sequential breaking of two phosphonate-Ca2+ bonds. We believe that this slow process of desorption of alendronate from the CaP coatings is primarily responsible for the slow release rates of alendronate. The other possibility for alendronate release would be via bulk dissolution of the CaP coating. However, this was not a significant factor in the in vitro release study because the Ca 2 + (1 .5x l0 " 3 M) in the buffer suppressed the solubility of CaP by the common ion effect. In fact, the CaP coating can be kept in the buffer for a few weeks without significant dissolution. During the release test, all alendronate concentrations in the buffer solution were determined to be at 10" 7 M level, except for the final sample made by dissolving the coating. It should be noted that, the concentration of alendronate was detectable by the HPLC method, but was below the quantitation limit. Most likely it may result in overestimation of alendronate concentrations because the calibration curve had a decreased slope (due to lower signal/noise ratio) at the lower end. So quantification in this region would give higher nominal reading than the actual concentration. Therefore, the release rate of alendronate from the coating shown in Fig 4.20 might be slightly overestimated. Nevertheless, this does not alter the fact that a slow release of alendronate over a prolonged period of time was achieved. 75 4.5 Conclusions In this study, a stable ELD process was developed to prepare uniform microporous CaP coatings on Ti and Ta. The coating solutions and deposition potentials were experimentally determined. Coating solution 3 (5.25 mM Ca, 10.5 mM P, pH 5) was found to be the optimum solution. ELD in solution 3 resulted in microporous coatings without DCPD and the coating deposition was relatively fast. The coating mainly contained OCP with a possible inclusion of apatite. With increase in deposition time, the coating changed from a cellular morphology into extended flake morphology; the crystal size and Ca/P ratio increased, and the crystalline disorder seemed to decrease. ELD current showed a rapid initial decrease followed by a prolonged nearly-constant stage following Cottrel's equation. Reduction of proton was the main electrode reaction at the first ELD stage, and electrolysis of molecular water became the main reaction with deposition time. Oxygen and nitrate were not critical for the ELD process. Alendronate was chemically adsorbed on the CaP coated Ti and Ta. The in vitro release of alendronate from porous Ta implants was prolonged, with only 10% eluted after 7 days. 76 Chapter 5 Electrolytic Deposition of Calcium Bisphosphonate Coatings 5.1 Introduction This chapter explores processing of a new type of bisphosphonate-containing coating on metallic implants. The objective is to load a higher dose of bisphosphonates than the surface adsorption, and also to achieve slow drug elution. The insoluble calcium salts of two bisphosphonate drugs (Ca-bisphosphonate), etidronate and alendronate, are electrolytically deposited as solid coatings on Ti and Ta. The morphology, composition, chemical structure and in vitro solubility of the coatings are characterized. In vitro release of alendronate from porous Ta implants is also studied. 5.2 Materials and Methods 5.2.1 Bisphosphonates Two bisphosphonates drugs were used in this study - etidronate and alendronate. Etidronate was purchased in the acid form [l-(hydroxyethylidene)-l,l-di(phosphonic acid), C2P20 7H 8] from Fluka. Alendronate was the same as used in the previous chapter (4.2.5). Their chemical structures are shown in Fig 5.1. OH C H , I I 3 I •Q(HO)OP — C — P O ( O H ) 0 " . N a + ( C H 2 ) 3 N H 3 + (HO) 2 OP — C — P O ( O H ) 2 OH (a) (b) Figure 5.1 Chemical structures of (a) etidronic acid and (b) alendronate; alendronate is a mono-sodium, zwitterionic salt. 5.2.2 Determination of Ca-bisphosphonates Precipitation Boundaries 2+ The precipitation boundaries of the bisphosphonates in the presence of Ca were determined and used for choosing ELD conditions. A series of 50 ml solutions containing the bisphosphonates and two-fold Ca(N03)2 were titrated with NaOH until the inception of turbidity; the precipitation boundaries of calcium bisphosphonates were correspondingly calculated. The ELD coating solutions for both etidronate and alendronate were chosen as 3.5 mM bisphosphonate and 7 mMCa 2 + . The pH of the ELD solutions was consistently set as 0.5 below the respective boundary pH, viz. 4.6 for etidronate and 4.8 for alendronate (for more details, referred to Results). 5.2.3 Electrolytic Deposition of Ca-Bisphosphonate Coatings Ti plates (20x20x2 mm) and porous Ta implants (diameter: 3.15 mm, length: 5.0 mm, Zimmer, Allendale, NJ, USA) were used as the substrates, and ELD was conducted in the three-electrode electrolysis mode. The sample pretreatment and ELD apparatus were the 78 same as described in 4.2.4. The ELD potentials established for Ti and Ta in the previous chapter were adopted. An electrode potential of -1.15 V (vs. SCE) was applied on Ti and -1.48 V was applied on porous Ta samples to form the respective calcium bisphosphonate coatings. After ELD, samples were rinsed with copious distilled water and air-dried. 5.2.4 Preparation of Reference Precipitates Because standard calcium salts of etidronate and alendronate were not commercially available, they were prepared in the lab and used for the subsequent characterization of the ELD coatings. NaOH solution (1 M) was added dropwise into 500 ml ELD solutions under stirring to the final pH of 7.6 (etidronate) or 7.8 (alendronate). Precipitation occurred immediately upon adding NaOH. The suspensions were stirred for 1 hour. The precipitates were separated by centrifugation (Galaxy Mini centrifuge, VWR) and purified by three cycles of dispersion in water and centrifugation. The collected precipitates were dried overnight at 60 °C. Because the chemical structure of bisphosphonates is well-known to be stable under the common aqueous conditions [141], the precipitates thus prepared can be used as standard substances to examine whether the ELD process altered the structure of the bisphosphonate drugs. In the following text, they will be described as reference precipitates. 79 5.2.5 Materials Characterizations 5.2.5.1 Morphology and Elemental Composition Samples were sputtered with Au-Pd alloy and studied with SEM for surface morphology. The elemental compositions were analyzed with the EDS spectrometer attached to the SEM. 5.2.5.2 Chemical Structure Infrared Spectroscopy To characterize the chemical structures of the ELD coatings, FTIR spectra of the coatings, reference precipitates and the starting etidronate and alendronate were collected and compared. More instrumentation details of SEM and FTIR were described in the previous chapter (referred to 4.2.5). Ion Chromatography 1 Ion chromatography (IC) was used to identify the chemical structures of the coatings. IC was chosen because of its ability to detect bisphosphonate drugs and impurities in them [142,143]. The ELD coatings were dissolved by diluted HC1 and analyzed with an ion 1 Electrospray ionization mass spectrometry (ESI-MS) was initially used for structure identifications. ESI-MS can precisely identify the molecular weight/charge ratio of main ionic species, but cannot detect low-level impurities. Therefore, it was later replaced by IC. 80 chromatography system (ICS 1000, Dionex, Sunnyvale, CA, USA) consisting of an IonPak AS 16 anion exchange analytical column, an ASUS Ultra II suppressor and a DS6 conductivity detector. In the IC system, anions were separated by the anion exchange column according to their different affinities to the column, and were subsequently detected by their conductivities. With this technique, each anion appeared as an IC peak at a unique elution time. The IC system worked under the standard conditions specified by the manufacturer (eluent: 35 mM NaOH, flow rate: 1 ml/min; solvent for eluent: deionized water, resistivity: 18.2 MQ.cm). ELD solutions and bisphosphonates solutions were tested as standards. The retention times of all other anionic species involved were determined by testing their respective standard solutions. 5.2.5.3 Coating Solubility To study the in vitro solubility of the coatings, 15 Ti samples ELD coated for 3 hours were each placed in a sealed polyethylene vial containing 5 ml "physiological" buffered solution (1.5 mM CaCl 2 , 140 mM NaCl, 20 mM HEPES, pH 7.40; refer to 4.2.6.5) and kept in a 37 °C water bath. Three samples were removed at regular intervals and the solutions were kept for analyses. The solubility of Ca-etidronate coating was determined by a spectrophotometric method [144]; all chemicals were ACS reagent grade and purchased from Sigma-Aldrich. 81 In brief, bisphosphonates released into the solution was oxidized into phosphate by potassium persulfate (K 2 S 2 0 8 ) in an autoclave (15 psig, 120 °C, 30 minutes). Phosphate reacted with ammonium molybdate [(NH4)6Mov024] to form a phosphomolybdate complex, which has a deep blue color. The complex was extracted with an organic phase (cyclohexane/methyl-isobutylketone, V:V=1:1), and the absorbance of the organic phase at 725 nm was measured with a UV-Vis spectrophotometer (Cary 50, Varian, Palo Alto, CA, USA). The calibration curve was established by measuring standard phosphate solutions, and the linear range was 1-5 X 10"5 M phosphate (R2>0.99, see Appendix B). Halfway through the project, the IC system (referred to 5.2.5.2) was purchased and used for studying the solubility of the Ca-alendronate coating. Under the standard IC conditions (5.2.5.2), the chloride (CI") peak (from NaCl in the buffer) partially overlapped with alendronate, interfering with the detection. The problem was solved by decreasing the NaOH eluent concentration to 28 mM. Other chromatography conditions were not changed. The calibration curve was established by measuring standard alendronate solutions, and the linear range was 2.5 - 30 x 10"6 M (R2 >0.99, peak area ~ concentration, see Appendix B). For the assay of both bisphosphonates, solutions were pre-tested and appropriately diluted, when necessary, to meet the linear ranges. 82 5.2.5.4 In vitro Release Three Ca-alendronate coated porous Ta implants were used to study the in vitro release. Each implant was placed in a microtube containing 2 ml of the "physiologic" buffer solution, capped and kept in a 37 °C water bath. 1 ml solution was withdrawn at regular times (i.e. 12, 24, 36, 48, 72, 96, 120, 144 and 168 hours) and kept in a sealed microtube for analyses. After each withdrawal, the microtube was replenished with 1 ml of the solution to keep the solution volume constant. After day 7, all specimens were removed, rinsed with copious water, air-dried and immersed in 1.5 ml HCI to dissolve any possible remaining coatings. After the dissolution procedure, Ta specimens were dried and examined under an optical microscope; no residue was observed. The 1.5 ml HCI solutions were also kept for analyses. Alendronate concentrations in the solutions were assayed by the IC method described in the prior section (5.2.5.3). 83 5.3 Results 5.3.1 Precipitation Boundary and E L D Solution The precipitation boundaries of etidronate and alendronate (Fig 5.2) increased with decreasing C a 2 + concentration, and the slope was sharper at lower than 2 m M . For both bisphosphonates, when their concentrations were below ~1.2 m M the precipitation became less obvious, and became difficult to judge at below 1 m M . Therefore, at lower concentrations, small differences in bisphosphonates concentration would cause relatively high difference in the pH at which the coating can form. This would adversely affect the robustness of the E L D processing. The slopes of the curves gradually decreased with increasing concentration; however, higher drug concentrations represent higher costs and lower yield, defined as the ratio of drugs deposited/dissolved. As a result, intermediate concentrations of 3.5 m M bisphosphonate and 7 m M C a 2 + were chosen for the E L D of both alendronate and etidronate. The pH of the E L D solutions was consistently chosen as 0.5 below the respective boundary pH, viz. 4.6 for etidronate and 4.8 for alendronate. The conditions of the E L D solutions are marked by arrows in Fig 5.2. The pH margin of 0.5 was experimentally determined, as the solutions were sufficiently close to the boundaries but remained stable for days without spontaneous precipitation. E L D solutions also contained 150 m M NaCl to buffer the ionic strength and to increase the conductivity. C a 2 + Concentration (mM) 4 8 12 -T 16 T 2 4 6 Etidronate Concentration (mM) (a) C a 2 + Concentration (mM) 0 4 8 12 16 I • 1 • 1 ' 1 ' r 4 I i i i i i i i i i 0 2 4 6 8 Alendronate Concentratin (mM) (b) Figure 5.2 Precipitation pH boundary of (a) etidronate and (b) alendronate in the presence of Ca 2 + , at fixed molar ratio of Ca:bisphosphonate = 2:1; error bars indicate SD (n=3). 85 5.3.2 Reference Precipitate The Ca-etidronate and Ca-alendronate reference precipitates were both nano-sized particles (Fig 5.3), most with a diameter smaller than 200 nm. The nano-sized particles further agglomerated into bigger "secondary" particles. The Ca/P ratio was measured by EDS to be 0.92 for Ca-etidronate and 0.63 for Ca-alendronate. Figure 5.3 SEM micrographs of (a) Ca-etidronate and (b) Ca-alendronate precipitate (preparation conditions see 5.2.4). 5.3.3 Coating Morphology and Composition For both etidronate and alendronate, after ELD for 15 minutes, coatings fully covered the Ti surface. The coatings presented rainbow-like color, indicating their thickness being sub-micrometer. The coatings displayed increasingly grayish color after ELD for 30 and 60 minutes, indicating growing thickness with time. Fig 5.4 and 5.5 show SEM morphologies of the coatings. The 15 minutes coatings were relatively smooth, slightly cracked and 86 composed of spherical particles about 500 nm in diameters. The 30 and 60 minutes coatings were rougher, extensively cracked and composed of densely packed globular domains (> 1 Jim), which were aggregated from the nano-sized primary particles. With further increase in ELD time, the coating morphologies did not change significantly, although the globular agglomeration became more pronounced and the density and width of cracks increased. The Ca/P ratios of the coatings did not change significantly with ELD time (Table 5.1), suggesting the coating compositions to be uniform. ELD of Ca-alendronate was also conducted on the porous Ta, and the coatings obtained were uniform and showed a similar morphology as on flat Ti (Fig 5.6). Table 5.1 Ca/P atomic ratio of Ca-bisphosphonate coatings prepared on Ti by ELD (-1.15 V vs SCE, 3.5 mM bisphosphonate, 7 mM Ca 2 + , pH 4.6 for etidronate and 4.8 for alendronate ) for different times (mean ±SD, n=3). ELD time (min) Ca-etidronate Ca-alendronate 15 0.46 ±0.01 0.25 ±0.01 30 0.45 ± 0.01 0.28 ± 0.01 60 0.45 ± 0.02 0.27 ± 0.02 87 (a) (b) <? JT « * JI , > '». % # • , 41, ^ 3 . ' -20 Ll 11T* • 1 * '* (c) (d) 1 > 20 Lim \ 5 (AIT) > • , , T ' 5 SE WD" 5..7mrii" 5. OOkV xICJi#' 5um / (e) (f) Figure 5.4 SEM micrographs of Ca-etidronate coatings on Ti plate for (a, b) 15 minutes, (c, d) 30 minutes and (e, f) 60 minutes (ELD conditions: -1.15 V vs SCE, 3.5 mM etidronate, 7 mM Ca 2 + , pH 4.6). The coatings built up from dense agglomeration of nano-sized particles. (a) (b) J \ , ' 20 j i m 1 \ ' * ; a4Jk % • ft V 5 j,im 3E WD 4.,9>JI, 5 . 0 0 W Ms : 20um • SE . ' WD 4.9mm 5:OQkV x lOlc : Sum (c) (d) 20 j.im i w 5 u.m SE: WD 7:7nun 5.OOkV x2 . Ok 20ura SE WD 7.7mra : 5.OOkV xlOk. 5um (e) (f) Figure 5.5 SEM micrographs of Ca-alendronate coatings on Ti plate for (a, b) 15 minutes, (c, d) 30 minutes and (e, f) 60 minutes. (ELD conditions: -1.15 V vs SCE, 3.5 mM alendronate, 7 mM Ca 2 + , pH 4.8). The coatings built up from dense agglomeration of nano-sized particles. 89 Figure 5.6 SEM micrographs showing a uniform Ca-alendronate coating on a porous Ta implant; (b) is the higher magnification view of the marked area in (a). (ELD conditions: - 1.48 V vs SCE, 3.5 mM alendronate, 7 mM Ca 2 + , pH 4.8, 30 minutes.) 5.3.4 Coat ing Chemica l Structure Figure 5.7 shows the ion chromatograms of Ca-etidronate and Ca-alendronate coatings prepared by 30 minutes of ELD on Ti. The peak 4 (6.8 min, Fig 5.7 a) and peak 5 (4.3 min, Fig 5.7 b) matched standard etidronate and alendronate, respectively. In addition, by testing standard samples, peak 2 was identified as chloride. Chloride appeared as strong peaks on the chromatograms of the ELD coatings because HC1 was used to dissolve the coatings. Chloride was also present in deionized water and bisphosphonates standards as trace impurities, therefore appearing as weak peaks. Peak 3 was identified as carbonate (C0 3 2"). The negative peak 1 on all chromatograms was water in the sample plugs, which had lower conductivity than the NaOH eluent. 90 Fig 5.8 displays the FTIR spectra of Ca-etidronate and Ca-alendronate coatings deposited for varying times on Ti. A l l coatings matched their respective reference precipitates. A strong band at 1100 cm"1 (etidronate, Fig 5.8 a) or 1090 cm"1 (alendronate, Fig 5.8 b) with two medium bands at 1000, 961 cm"1 (etidronate) or 1000, 964 cm"1 (alendronate) are characteristic of the P-O stretching modes of phosphonate group [145]. Two medium bands at 567, 490 (etidronate) or 566, 497 cm"1 (alendronate) are characteristic of the bending modes of phosphonate group [145]. 91 1 .0 > '1 0.0 TJ cz o O CO ;> o T3 C o O -0.5 ELD Coating Etidronate L Water 23 2 4 6 Time (minutes) (a) 2 4 6 Time (minutes) (b) Figure 5.7 Ion chromatograms of (a) Ca-etidronate, (b) Ca-alendronate ELD coatings on Ti and their respective standards; Peak 1: water of sample plug, 2: CI", 3: CO3 2 " . (ELD conditions: -1.15 V vs. SCE, 30 minutes, 3.5 mM bisphosphonate, 7 mM Ca 2 + , pH 4.6 for etidronate and 4.8 for alendronate.) 92 1200 800 400 Wavenumber (cm1) (a) \ A Ca alendronate A / ' / \ W \ Ref precipitate / V — . ^A^JO min ^ ^ / v ^ / ^ ^ ~ N ^ V 30 min ^ _J^ ^\ 1 1200 800 40 Wavenumber (cm1) (b) Figure 5.8 FTIR spectra of (a) Ca-etidronate and (b) Ca-alendronate ELD coatings on Ti and their respective reference precipitates. (ELD conditions: -1.15 V vs. SCE, 3.5 mM bisphosphonate, 7 mM Ca 2 + , pH 4.6 for etidronate and 4.8 for alendronate.) 93 5.3.5 In vitro Solubility In vitro solubilities of the coatings are shown in Fig 5.9. In the case of Ca-etidronate coating, the drug concentration was 8 xlO"5 M at day 1, and then it decreased to 6.8><10"5 M at day 2 and remained relatively stable at 6 xlO"5 M up to day 8. Starting from day 4, some solid particles formed in the solution. The particles were filtered off, since the target of the solubility measurement was the free etidronate concentration. In the case of Ca-alendronate coating, the concentration was 1.5 x i o - 4 M at day 1; it reached 2.5xl0"4 M at day 4 and remained relatively stable up to day 8. A l l coatings studied survived 8 days of soaking without complete dissolution. 5 ^ 4 E 3 c o * CD 2 -t—' c CU o c 1 o ' o 0 0 2 4 6 8 Time (days) Figure 5.9 In vitro solubility of Ca-bisphosphonate ELD coatings on Ti: concentration in the buffer solution with time; error bars indicate SD (n=3). (ELD conditions: -1.15 V, 3h, 3.5 mM bisphosphonate, 7 mM Ca 2 + , pH 4.6 for etidronate and 4.8 for alendronate) 94 5.3.6 In vitro Release Fig 5.10 shows the in vitro release of alendronate from porous Ta implants with Ca-alendronate coatings (ELD for 30 minutes). Alendronate release was fast in the first day (i.e. 94 jig, or 68 %) but gradually slowed down; all alendronate was released within 3 days. The total dose of alendronate loaded by the coating was 137 ±10 ug. 160 h Time (days) Figure 5.10 In vitro cumulative release of alendronate from porous Ta implants ELD coated with Ca-alendronate coatings for 30 minutes; error bars indicate SD (n=3). (ELD conditions: -1.48 V, 30 minutes, 3.5 mM bisphosphonate, 7 mM Ca 2 + , pH 4.6 for etidronate and 4.8 for alendronate) 95 5.4 Discussion In this study, the ELD technique developed in the previous chapter was successfully extended to deposit bisphosphonate drugs as solid coatings on Ti and Ta. To our knowledge, this is the first study that reports the deposition of bisphosphonates solid coatings for local drug delivery. Bisphosphonates, as organic analogues of pyrophosphate, are thermally unstable (decompose at < 2 0 0 °C) and are not suitable for most coating techniques (e.g. thermal spray, sol-gel). In this study, this problem was avoided by using the room temperature ELD technique. The process was similar to the ELD of calcium phosphate coatings: the local pH jump caused by the current deprotonated the drugs into the anionic forms; the drug anions combined with Ca 2 + to form insoluble salts. Coatings of both bisphosphonates showed hierarchical agglomeration of nano-sized particles. The agglomeration from the nano-sized particles may be driven by the pH change (i.e. gelation). The morphologies of the coatings are consistent with the reference precipitates (i.e. nano-particles), but the Ca/P ratios of the coatings are lower than the precipitates. This may be best explained by the fact that the pH in the ELD conditions was lower than the ones in the preparation of reference precipitates. This is reminiscent of the ELD of calcium phosphate coatings at the same electrode potentials, where the pH only partially deprotonated phosphate species to form OCP rather than pure H A [refer to Chapter 4 ] . The coatings were extensively cracked, mainly due to 96 the significant shrinkage during the drying process. A few trials showed that if the coated samples were dried in supercritical CO2, the crack density was reduced, but not eliminated. A major concern in ELD of organic molecules is whether they are electrochemically altered (e.g. reduced) at the electrodes. In this study, the chemical structures of the coatings were confirmed by IC and FTIR analyses. On ion chromatograms of the coatings, the respective bisphosphonates were detected, without appreciable impurities. This strongly indicated that the drugs were not altered. Because bisphosphonates had multiple negative charges, any changes in structure would form new anionic species and be detected by the IC. In addition, the FTIR patterns of the ELD coating were nearly identical with the reference precipitates. This excluded the possibility that the drugs were altered into another anion but with close IC retention time. Therefore, the ELD coating was confirmed to be solely calcium salt of the respective bisphosphonates. The rationales of developing Ca-bisphosphonate solid coatings are: 1) high dose of drugs could be deposited and 2) the drug concentration can be limited by the low solubility. These hypotheses are supported by the results. The dose of the ELD coatings depends on the coating thickness and therefore ELD time. On the porous Ta implants, 30 minutes of ELD loaded ~ 140 jig of alendronate at ~500 nm coating thickness. By extending the ELD time, 2-3 (im thick coatings were easily fabricated. In the in vitro solubility study, Ca-etidronate equilibrated with the buffer solution at 6 xl0~5 M within 4 days, and Ca-97 alendronate equilibrated at 2.5><10"4 M also within 4 days. In the case of Ca-etidronate, solid particles formed in the solution after 4 days, indicating the drug salt transformed from amorphous nano-particles into a crystalline form, as governed by solubility principles. In the in vitro release of Ca-alendronate (Fig 5 .10) , the drug concentration was at 9 x l O " 5 M level for the first 1.5 days, and declined to 4 x 1 0 " 5 M after 3 days. The concentrations were significantly lower than the solubility level, because insoluble salts usually dissolve slowly, as evidenced by the time (> 2 days) needed to reach the equilibrium level. Bisphosphonates have complex effects on bone cells. In vitro studies showed that etidronate inhibited osteoclast at 10~6 M level but did not affect the osteoblast viability up to 1 mM [ 1 4 6 ] , and alendronate effectively inhibit osteoclasts at 10" 9 M level but decrease the osteoblast viability beyond 1 X 10" 4 M [147 , 148] . Thus, the solubility value suggests Ca-etidronate coating may reduce bone resorption without impairing bone formation, and Ca-alendronate coating may potently reduce bone resorption but might also affect bone formation. The in vitro release results, however, suggest the free alendronate concentration to be safe to osteoblasts. Due to its high clinical interests, the effect of Ca-alendronate coating will be investigated in an in vivo model in the next chapter. 98 5.5 Conclusions Continuous coatings of calcium etidronate and alendronate were ELD deposited at room temperature on flat Ti substrates and porous Ta implants. The technical steps and mechanisms involved are the same as ELD of calcium phosphate coatings. The chemical structures of the bisphosphonates are not changed by the electrolytic process. The solubilities of the coatings in the buffer solution are of 6 x l O " 5 M for Ca-etidronate and 2.5 x l O " 4 M for Ca-alendronate. In vitro study showed that drug release from the Ca-alendronate coated porous Ta implants was completed within 3 days and the alendronate concentration was below the solubility limit. The coatings are expected to provide local release of high dose of bisphosphonate drugs, from metallic implants. 99 Chapter 6 In vivo Study of Bisphosphonate-containing Coatings 6.1 Introduction Faster implant fixation and less loosening are important to the performance of artificial joint implants. A more positive bone modeling has been proposed as a key to achieve both goals, and this may be realized through local delivery of bisphosphonates from the implant. In the preceding chapters, two types of alendronate-containing coatings have been developed on porous Ta implants. In this chapter, the performance of the two types of coatings was studied using porous Ta implants in an animal model. The objective was to study the affects of the coatings on the implant fixation and new bone formation. 100 6.2 Materials and Methods 6.2.1 Implants Four types of cylindrical porous Ta implants (diameter: 3.15 mm, length: 5.0 mm) were used in this study. They include: Two control groups: 1. As-received porous Ta implants. In the following text, these implants will be abbreviated as Ta. 2. Porous Ta implants ELD coated with CaP, as described in Chapter 4. These implants will be abbreviated as Ta-CaP. Two implant groups with ALN-containing coatings: 3. Porous Ta implants ELD coated with CaP and chemically adsorbed with 10.7 |lg alendronate, as described in 4.2.4. The coating will be abbreviated as CaP-A L N and these implants will be abbreviated as Ta-CaP-ALN. 4. Porous Ta implants ELD coated with Ca-Alendronate for 30 minutes to load 137 |_Lg alendronate, as described in 5.2.3. The coating will be abbreviated as CaALN, and these implants will be abbreviated as Ta-CaALN. The alendronate dose of this coating (i.e. 30 min ELD) was 12 times of the dose of group 3, and the difference was considered significant. It is therefore chosen to evaluate the effect of the CaALN coatings. 101 6.2.2 Animals Thirty three adult, female New Zealand white rabbits, weighting between 3.8 to 5.5 Kg were used to evaluate the effect of the coatings on implant fixation and new bone formation. The rabbits were randomly divided into four groups for the four sample types, as listed in Table 6.1. Table 6.1 Number of rabbit used and porous Ta implant tested. Implant Group Implantation Time (weeks) Number of Rabbits Number of Implants for Push-out Number of Implants for Histomorphometry Ta 4 8 5 5 5 6 5 6 Ta-CaP 4 4 7 7 Ta-CaP-ALN 4 8 5 5 7 7 7 7 Ta-CaALN 4 8 5 4 7 6 7 6 102 6.2.3 Implantation Procedure Two implants were placed unicortically in the right anteromedial tibia. The two implants were 10 mm apart and the proximal one was about 30 mm away from the proximal end of the tibia (Fig 6.1 a-c). To prevent potential cross-influence of alendronate, the two implants placed in the same rabbit were of the same type. The surgery was performed under anesthesia by intramuscular injection of ketamine-xylazine and continuous inhalation of isoflurane. A small longitudinal incision was made on the tibia and the periosteum was reflected. A 5/64 inch (1.98 mm) hole was drilled perpendicular to the bone facet and through the cortex, using a.low speed handset drill (Minidriver, Linvatec, Largo, FL, USA). The hole was enlarged with a 1/8 inch (3.17 mm) twist drill bit. Both drillings were performed under saline irrigation to prevent overheating of the adjacent bone. The implant was gently inserted and pushed with a custom-made stainless steel punch with 1 mm recess to make the length of the implant protruding the cortex consistent. A l l implants and drill bits were beta-ray sterilized before implantation. Fig 6.1 b and c show X-ray radiographs of two implants placed in a tibia. Al l rabbits were allowed to bear body weight after surgery and had free access to water and food. The rabbits were sacrificed at 4 and 8 weeks post-operatively by injection of pentobarbital. The animal study protocol was approved by the Animal Care Committee of the University of British Columbia (Appendix C). 103 (a) (b) (c) Figure 6.1 (a) Photograph showing two porous Ta implants placed in a rabbit tibia, (b-c) Two postmortal radiographs showing the top and side view of two porous Ta implants placed in a rabbit tibia. 104 6.2.4 Push-out Test The bone-implant fixation was evaluated by push-out tests. The tibiae were harvested and stripped of soft tissues. Before each tibia was tested, both the proximal and the distal portions were manually cut off with a low speed diamond blade saw at about 10 mm away from the nearest Ta implant. A longitudinal cut was made through the center of the bone to separate the posterolateral half from the anteromedial half of the tibia. The anteromedial half of the tibia (with two implants) was rinsed with water to remove the bone marrow. Fig 6.2 shows the photographs of a prepared specimen. The processed bone/implant specimen was tested with a Minimat 2000 mechanical testing system (Rheometric Scientific, Piscataway, NJ, USA) (Fig 6.3). The specimen was supported on the external bone surface by a steel washer with an inner diameter of 5.5 mm. The space between the curved bone surface and the steel washer was filled with cured P M M A resin to facilitate an even load distribution. The endosteal end of the Ta implant was loaded through a 5 mm tungsten carbide sphere and the implant was pushed out at the crosshead speed of 0.5 mm/minute. The force-displacement curve was monitored by a computer and the test were stopped shortly after the peak load was passed (i.e. failure occurred). The peak value on the force-displacement curve was the push-out failure force. Throughout the test, the specimen was kept wet with Kimwipe paper tissue soaked with phosphate buffered saline solution (1.06 mM K H 2 P 0 4 , 5.60 mM Na 2 HP0 4 , 0.14 M NaCl, pH 7.4). 105 The push-out shear strength (f) was calculated based on the failure force (F) and the thickness (L) of the host bone cortex: T= F/(7tDX L), where D is the implant diameter (3.15 mm). The cortex thickness (L) was measured on the ground sections of the embedded samples (see the following section: Histomorphometry), using a calibrated optical microscope. Three measurements were made on the distal, medial and anterior sides of the implant, and the mathematical average was used as the L. The push-out strength thus computed is a nominal strength, which may be higher than the real interfacial shear strength as the thickness of the periosteal and endoseal calluses were not considered. However, it is valuable from the clinical perspective, since it indicates how strong the implant is anchored in the implant bed for a given cortex thickness. (a) (b) Figure 6.2 Photographs of a bone cortex specimen with a porous Ta implant prepared for the push-out test; (a) endosteal view and (b) side view. 106 force washer (a) (b) Figure 6.3 (a) Schematic and (b) photograph of the push-out test; the specimen is supported on a steel washer and the space between the implant and the P M M A resin is filled with cured PMMA; the implant is push from the endocortical side at 0.5 mm/minute. 107 6.2.5 Histomorphometry After push-out tests, the samples were fixed with 10% formalin (Fisher), sequentially dehydrated in an ascending series of acetone/water solutions, infiltrated and embedded in epoxy resin (Spurr, Polysciences, PA, USA). The embedded samples were ground perpendicular to the bone axis from the distal side. During grinding, the exposed implant was frequently examined with an optical microscope (Nikon Epiphot 300) to measure the cortex thickness and fine-adjusted to make sure the final section was parallel to the long axis of the implant and cross its center. At least three measurements were made on each sample, and the mean value was used as the cortex thickness. The ground samples were polished with 6 um and 1 urn diamond slurry and finally vibration polished with 0.05 um silica for 30 minutes. The polished samples were sputter-coated with Au/Pt alloy and studied with SEM in backscattered electron (BSE) imaging mode. The grey scale contrast, created by the difference in average atomic numbers (e.g. difference in CaP mineral content), made it possible to distinguish the new bone from the host bone, Ta, soft tissues and epoxy resin (Fig 6.4). Then new bone within and outside the host cortex defect area were recognized by shifting the selected areas (Photoshop Element, Adobe, San Jose, CA, USA) and termed ingrowth and callus, respectively. The area of the new bone, Ta and soft tissue and epoxy resin were measured (in mm2) by an image analysis software (Clemex VisionPro, Clemex, Longueuil, QC, Canada). The ingrowth bone was also expressed as ingrowth percentage: Ingrowth percentage = ingrown bone area (mm2)/ cortex defect area (mmz) 108 Figure 6.4 Procedure of histomorphometric analyses, (a): In the BSE micrograph, the host bone and new bone showed different greyscales; (b): By image analysis (greyscale thresholding), the host bone cortex was replaced with yellow, the new bone was replaced with blue, and the Ta was replaced with white. The area of each color was then measured. 109 6.2.6 Statistical Analyses Statistical analyses were performed by A N O V A followed by Fisher's PLSD (protected least significant difference) test using SPSS 12.0 statistical package (SPSS, Chicago, IL, USA) and p<0.05 was considered as statistically significant. 6.3 Results The results of push-out test and histomorphometry are listed in Table 6.2. The statistical analyses results (p values) are listed in Table 6.3 - 6.6. 6.3.1 Push-out Force and Strength General Observations On the push-out force-displacement curves, many 4 week implants displayed a nearly linear increase of force with displacement until a peak value, followed by a short non-linear (yield) stage and finally a sudden force drop indicating the failure (Fig 6.5). Eight week samples usually displayed a longer yield stage with increase in force after the initial linear stage. The last stage of force increase was due to the stronger bone-implant interface. At both week 4 and week 8, failure occurred predominantly at the interface between the host bone and the new bone, with only occasional fracture in the host bone cortex. no 600 Failure "0.0 0.5 1.0 1.5 2.0 2.5 Displacement (mm) Figure 6.5 Representative push-out curves of two porous Ta implants after implantation for 4 and 8 weeks respectively; the 8 week specimen displayed an increase in force after the deformation of porous Ta. Week 4 Figure 6.6 a shows the push-out force of the implants. Four weeks after implantation, the mean push-out force was 236 ± 32 N for Ta, 285 ± 53 N for Ta-CaP, 334 ± 82 N for Ta-CaP-ALN and 229 ± 52 N for Ta-CaALN. The force of Ta-CaP was 20% higher than Ta (p=0.177, Table 6.3), and Ta-CaP-ALN was 17% further higher than Ta-CaP (p=0.134), but both differences were not statistically significant. However, when the comparison was made between Ta-CaP-ALN and Ta, the difference was 41% and statistically significant (p=0.010). The force of Ta-CaALN was 3.1% lower than Ta (p=0.837) and 19.6% lower than Ta-CaP (p=0.093), but both differences were not statistically significant. 111 Fig 6.6 b shows the push-out strength of the implants. Four weeks after implantation, the mean push-out strength was 21.4 ± 3.4 MPa for Ta, 23.8 ± 4.5 MPa for Ta-CaP, 28.3 ± 5.4 MPa for Ta-CaP-ALN and 20.3 ± 4.0 MPa for Ta-CaALN. Therefore, the strength of Ta-CaP was 11% higher than Ta (p=0.384, Table 6.3), and Ta-CaP-ALN was further 21% higher than Ta-CaP (p=0.072), but both not statistically significant. However, when the comparison was made between Ta-CaP-ALN and Ta, the difference was 32% and statistically significant (p=0.016). The strength of Ta-CaALN was 5.6% lower than Ta (p=0.658) and 13% lower than Ta-CaP (p=0.157), but both differences were not statistically significant. Week 8 Eight weeks after implantation (Fig 6.6 a), the push-out force was 411 ± 52 N for Ta, 411 ± 80 N for Ta-CaP-ALN and 333 ± 60 N for Ta-CaALN, showing 74%, 23% and 45% increase from week 4, respectively. Ta-CaALN was 19% lower than Ta and Ta-CaP-ALN, but no statistically significant differences were found by ANOVA. The push-out strength (Fig 6.6 b) was 31.5 ± 4.1 MPa for Ta, 34.4 ± 4.4 MPa for Ta-CaP-ALN and 28.4 ± 4.2 MPa for Ta-CaALN, showing 47%, 22% and 40% increase from week 4, respectively. Ta-CaP-ALN was 9.3% higher than Ta and Ta-CaALN was 9.7% lower than Ta, but no statistically significant differences were found by ANOVA. Table 6.2 Summary of push-out and histomorphometric results. Ingrowth Ingrowth Callus Area Total New Bone Push-out Push-out Time + Implant n' 2 , 2, (week) Percentage Area (mm) (mm) Area (mm') Force (N) Strength (MPa) Ta 5 0.311±0.15 1.18±0.50 1.30±O.5O 2.48±0.88 236±32 6 0.466±0.035 1.99±0.45 2.00± 1.07 3.99±1.46 411±52 21.4±3.4 31.5±4.1 Ta-CaP 7 0.291±0.063 1.28±0.15 1.80±0.35 3.08±0.48 285±53 23.8±4.5 Ta-CaP-ALN 7 0.386±0.063 1.57±0.19 2.67±0.89 4.24±0.93 334±82 7 0.367±0.039 1.61±0.14 1.90±0.78 3.84±0.87 411±80 28.3±5.4 34.4±4.4 Ta-CaALN 7 0.198±0.075 0.78±0.26 1.81±1.03 2.58±1.20 229±52 6 0.408±0.082 1.62±0.45 2.22±081 3.53±0.: 333±60 20.3±4.0 28.4±4.2 * The number of porous Ta implants tested. 600 S 400 Z J O sz CO 13 D_ 200 0 ITa t ^ T a - C a P r~~1 Ta-CaP-ALN • I Ta-CaALN * Week 4 Week 8 (a) Week 4 Week 8 (b) Figure 6.6 (a) Push-out force and (b) push-out strength of different porous Ta implants after implantation of 4 and 8 weeks; * indicates difference significant over Ta, and ** indicates difference significant over Ta-CaP. Statistically significant differences were found in 4 week implants; p values and number of implants (n) are listed in Table 6.3 (see next page). No statistically significant differences were found by ANOVA for 8 week implants; post hoc analysis was not adopted. 114 Table 6.3 List of p values of statistical analyses (ANOVA followed by PLSD test) for push-out force and strength of porous Ta implants after 4 weeks of implantation. Push-out force Ta (n=5) Ta-CaP (n=7) Ta-CaP-ALN (n=7) Ta-CaALN (n=7) Ta 0.177 0.010 0.837 Ta-CaP 0.134 0.093 Ta-CaP-ALN 0.003 Push-out strength Ta (n=5) Ta-CaP (n=7) Ta-CaP-ALN (n=7) Ta-CaALN (n=7) Ta 0.384 0.016 0.658 Ta-CaP 0.072 0.157 Ta-CaP-ALN 0.003 115 6.3.2 Histomorphometry 6.3.2.1 Week 4 Four weeks after implantation, new bone has generally grown into 2 or more layers of pores of the implant (Fig 6.7). The new bone inside the superficial pores appeared to be denser than the one inside the deeper pores. Ingrowth The ingrowth percentage (Fig 6.8 a) was 0.311 ± 0.150 for Ta, 0.291 ± 0.063 for Ta-CaP, 0.386 ± 0.063 for Ta-CaP-ALN, and 0.198 ± 0.075 for Ta-CaALN. The ingrowth bone area (Fig 6.8 b) was 1.18 ± 0.50 mm2 for Ta, 1.28 ± 0.15 mm2 for Ta-CaP, 1.57 ± 0.19 mm2 for Ta-CaP-ALN and 0.78 ± 0.26 mm2 for Ta-CaALN. Compared with Ta, Ta-CaP had 6.3% lower ingrowth percentage (p=0.704, Table 6.4) but 8.6% higher ingrowth area (p=0.568), but both were not significant. The seeming discrepancy between the percentage decrease and the area increase was attributed to the mean cortex thickness of Ta-CaP group being greater than Ta group by 8.9% (but without statistically significant difference). Ta-CaP-ALN displayed 24% higher ingrowth percentage (p=0.164) and 33% higher ingrowth area (p=0.029) than Ta; the difference in ingrowth area was statistically 116 significant. Ta-CaP-ALN had 32% higher ingrowth percentage (p=0.058) and 23% higher ingrowth area (p=0.067) than Ta-CaP, but both differences were not statistically significant. Ta-CaALN showed 36% lower ingrowth percentage (p=0.038) and 34% lower ingrowth area (p=0.024) than Ta, both differences were statistically significant. When compared to Ta-CaP, Ta-CaALN showed 32% lower ingrowth percentage (p=0.059) and 39% lower ingrowth area (p=0.003), and the difference in ingrowth area was statistically significant. Callus The callus area (Fig 6.9) was 1.30 ± 0.50 mm2 for Ta, 1.80 ± 0.35 mm2 for Ta-CaP, 2.67 ± 0.89 mm2 for Ta-CaP-ALN and 1.81 ± 1.03 mm2 for Ta-CaALN. Ta-CaP-ALN showed a remarkably higher callus area: 105% higher than Ta (p=0.005, Table 6.5) and 48% higher than Ta-CaP (p=0.043), both statistically significant. Other differences among implants were not significant: Ta-CaP 38% higher than Ta (p=0.272), Ta-CaALN 39% higher than Ta (p=0.266) and 0.3% higher than Ta-CaP (p=0.987). Total New Bone Summing up the ingrowth and the callus, the total new bone area (Fig 6.10) was 2.48 ± 0.88 mm2 for Ta, 3.08 ± 0.48 mm2 for Ta-CaP, 4.24 ± 0.93 mm2 for Ta-CaP-ALN and 2.58 ± 1.20 mm 2 for Ta-CaALN. The total new bone area of Ta-CaP-ALN was 71% higher than Ta (p=0.003, Table 6.6) and 38% higher than Ta-CaP (p=0.026), and both differences were statistically significant. Other differences among implants were not statistically significant: Ta-CaP 24% higher than Ta (p=0.275), Ta-CaALN 4% than Ta (p=0.847) and 16% lower than Ta-CaP (p= 0.323). (a) (b) (c) (d) Figure 6.7 BSE micrographs of different implants after implantation for 4 weeks; (a) Ta, (b)Ta-CaP, (c) Ta-CaP-ALN, (d) Ta-CaALN. The new bone appeared porous, and many existed in isolated island forms, especially at the central pores. 119 100 Week 4 (a) Week 8 Week 4 Week 8 (b) Figure 6.8 (a) Bone ingrowth percentage and (b) ingrowth area of different implants after implantation for 4 and 8 weeks; * indicates difference significant over Ta, and ** indicates difference significant over Ta-CaP. The p values and number of implants (n) are listed in Table 6.4 (see next page). No statistically significant differences were found by A N O V A for the ingrowth area in 8 week implants; post hoc analysis was therefore not adopted. 120 Table 6.4 List of p values of statistical analyses (PLSD test after ANOVA) for ingrowth percentage and ingrowth bone area of porous Ta implants. Week 4: Ingrowth percentage Ta (n=5) Ta-CaP (n=7) Ta-CaP-ALN fn=7) Ta-CaALN (n=7) Ta 0.704 0.164 0.038 Ta-CaP 0.058 0.059 Ta-CaP-ALN 0.001 Week 8: Ingrowth percentage Ta Ta-CaP-ALN Ta-CaALN (n=6) (n=7) (n=6) Ta 0.005 0.088 Ta-CaP-ALN 0.204 Week 4: Ingrowth area Ta (n=5) Ta-CaP (n=7) Ta-CaP-ALN (n=7) Ta-CaALN (n=7) Ta 0.568 0.029 0.024 Ta-CaP 0.067 0.003 Ta-CaP-ALN 0.000 121 6 I ITa ^ T a - C a P Week 4 Week 8 Figure 6.9 Callus area of different implants after implantation for 4 and 8 weeks; * indicates difference significant over Ta, and ** indicates difference significant over Ta-CaP. The p values and number of implant (n) are listed in Table 6.5 below. No statistically significant differences were found by A N O V A for 8 week implants; post hoc analysis'was therefore not adopted. Table 6.5 List of p values of statistical analyses (PLSD test after ANOVA) for callus area of porous Ta implants after implantation for 4 weeks. Ta (n=5) Ta-CaP (n=7) Ta-CaP-ALN (n=7) Ta-CaALN (n=7) Ta 0.272 0.005 0.266 Ta-CaP 0.043 0.987 Ta-CaP-ALN 0.044 122 Week 4 Week 8 Figure 6.10 Total new bone area of different implants after implantation for 4 and 8 weeks; * indicates difference significant over Ta, and ** indicates difference significant over Ta-CaP. The p values and number of implants (n) are listed in Table 6.6 below. No statistically significant differences were found for 8 week implants by ANOVA; post hoc analysis therefore was not adopted. Table 6.6 List of p values of statistical analyses (PLSD test after ANOVA) for total new bone area of porous Ta implants after implantation for 4 weeks. Ta (n=5) Ta-CaP (n=7) • Ta-CaP-ALN (n=7) Ta-CaALN (n=7) Ta 0.275 0.003 0.874 Ta-CaP 0.026 0.323 Ta-CaP-ALN 0.003 123 6.3.2.2 Week 8 Eight weeks after implantation, Ta and Ta-CaALN showed increase in all of the histomorphometric parameters compared with 4 weeks, whereas Ta-CaP-ALN showed decrease in nearly all of the parameters. For Ta-CaP-ALN implants (Fig 6.11), the bone inside the superficial pores became denser than week 4, but many isolated bone islands inside deeper pores disappeared. Ingrowth The ingrowth percentage of Ta (Fig 6.8 a) increased (by 50% from week 4) to 0.466 ± 0.035, Ta-CaALN increased (by 106% from week 4) to 0.408 ± 0.082, but Ta-CaP-ALN slightly decreased (by 5% from week 4) to 0.367 ± 0.039. Ta-CaP-ALN became 10% lower than Ta, and the difference was statistically significant (p=0.005, Table 6.4). Ta-CaALN was 12% lower than Ta, but without significant difference (p=0.088). The ingrowth area (Fig 6.8 b) of Ta increased (by 69% from week 4) to 1.99 ± 0.45 mm2, Ta-CaALN increased (by 109% from week 4) to 1.62 ± 0.45 mm2 and Ta-CaP-ALN marginally increased (by 3% from week 4) to 1.61 ± 0.14 mm2. The marginal increase of Ta-CaP-ALN was mainly attributed to the fact that the mean bone cortex thickness was 2% greater than week 4. Ta-CaP-ALN and Ta-CaALN were 19% lower than Ta, but no statistically significant differences were found by ANOVA. 124 Callus The callus area (Fig 6.9) of Ta increased (by 54% from week 4) to 2.00 ± 1.07 mm2, Ta-CaALN increased (by 5.4% from week 4) to 1.90 ± 0.78 mm2, but Ta-CaP-ALN decreased (by 17% from week 4) to 2.22 ± 0.81 mm2. Ta-CaP-ALN was 11% higher than Ta and Ta-CaALN was 5% lower than Ta, but no statistical significance differences were found by ANOVA. Total New Bone Summing up the ingrowth and the callus, the total new bone area (Fig 6.10) of Ta increased (by 61% from week 4) to 3.99 ± 1.46 mm2, Ta-CaALN increased (by 37% from week 4) to 3.53 ± 0.88 mm2, whereas Ta-CaP-ALN decreased (by 9.4% from week 4) to 3.84 ± 0.87 mm2. Ta-CaP-ALN was 3.8% lower than Ta and Ta-CaALN was 12% lower than Ta, but no statistically significant differences were found by ANOVA. 125 Figure 6.11 BSE micrographs of different implants implanted for 8 weeks; (a) Ta, (b) Ta-CaP-ALN and (c) Ta-CaALN. The new bone appeared to be denser than week 4, especially at the superficial pores. 126 6.4 Discussion 6.4.1 Main Findings The main findings of this study were 1) the CaP-ALN coating significantly enhanced the early (week 4) fixation of the porous Ta to the cortical bone; 2) at the later stage (week 8), differences between implants were not significant (Fig 6.6). The push-out strength of Ta-CaP-ALN implants at week 4 was 32% higher than Ta control; in fact, it reached 90% of the strength of Ta at week 8. Therefore, the improvement found in this animal model provides clinical potentials of accelerating implant fixation and reducing early micromotion, which may translate into shorter healing time and longer implant service life. Ta-CaALN implants showed similar push-out strength with Ta itself at both week 4 and week 8. As a short-term experiment, this study was unable to demonstrate its ability to inhibit long-term osteolysis. However, this pilot study demonstrated that the CaALN coating was safe and did not impair the implant fixation. In addition, because the CaALN coating can locally deliver high dose of A L N , it provides the possibility of preventing future osteolysis and implant loosening, which warrants further studies. This study also showed that different coatings affected implant fixation through their own influence on new bone formation. This is further discussed in the following section. 6.4.2 New Bone Formation and Implant Fixation Biological fixation of porous implants is achieved by interlock with the ingrown tissue. It is reasonable to hypothesize that larger bone mass associated with the porous implant will result in a stronger bone-implant fixation. When results from all implants are plotted together (Fig 6.12), push-out force shows a reasonably clear trend of increase with bone area. At week 4, the force-ingrowth area correlation shows a linearity (R2) of 0.43 and the force-total new bone area correlation shows a linearity of 0.51. At week 8, they have linearity of 0.33 and 0.46, respectively. The correlations support the hypothesis. Week 4 Compared with Ta, Ta-CaP showed 6% lower ingrowth percentage but 11% higher push-out strength. Therefore, the callus was the contributor to the higher strength of Ta-CaP than Ta. Indeed, after taking into account the callus area, the total new bone area of Ta-CaP was 21% higher than Ta. It may be mentioned that for all implants the callus showed similar greyscale with the new ingrowth bone, indicating a similar mineral content. Ta-CaP-ALN showed the highest ingrowth percentage, ingrowth bone area, callus area and total new bone area; the push-out strength was correspondingly the highest among all types of implants. In addition to the effect of CaP coating, the increases may be attributed to the capability of bisphosphonates to inhibit osteoclastic activities. The free A L N concentration of 10"7 M level in the in vitro release study (4.3.8) appears to be effective in inhibiting osteoclastic activities according to an in vitro study [148]. The 128 reasoning is also supported by two closely related studies in literature [149, 150]. Studies on rats and sheep found that during the healing of long bone fracture, osteoclastic activities began at as early as 2 weeks after the fracture, indicating that the callus already underwent resorption and remodelling in the early phase of bone healing. Therefore, it may be reasonable to deduce that in this study the locally delivered alendronate disrupted the osteoclastic activity to produce more bone and/or callus. A few studies of systemic bisphosphonates treatment found that bisphosphonates significantly increased the callus size during the long bone fracture healing and improved the strength of the healing bone [151, 152, 153]. In addition, L i et al observed a decrease in osteoclast numbers in the callus area in bisphosphonate treated animals, directly supporting the anti-catabolic mechanism [153]. In this study, the increase in total new bone area, and particularly the large increase in callus area highly resemble the effect of the systemic treatments and strongly indicate an inhibited osteoclastic activity. It is also interesting to note that Im et al recently found that alendronate increased the proliferation of primary human trabecular bone cells and an osteoblast-like cell line at 10"6-10"9 M level, attaining the peak at 10"8 M , and also increased osteoblastic differentiation of these cells [154]. These, anabolic effects of alendronate might also have contributed to the increase in new bone area observed in this study. In addition, we recently studied porous Ta implanted in rabbit trabecular bone and labeled the new.bone formation by periodical injection of fluorescent markers [155]. We found that for the Ta and Ta-CaP implants, new bone formation predominantly started in the center of the pores and subsequently expanded toward, but with minimal contact with, the surface of the struts. In contrast, for Ta-CaP-129 A L N implants, new bone formed predominantly on the surface of the Ta struts and subsequently grew inwardly to the center of the pores. These different patterns of new bone formation suggest an enhanced osteoblast attachment and/or osteoblastic activity by the CaP-ALN coating, seemingly in line with the anabolic effects observed in vitro [154]. However, the explanations from the perspective of osteoblasts need to be confirmed by more extensive studies and their relative role (vs. inhibition of osteoclasts) remains to be evaluated. Ta-CaALN showed lower ingrowth percentage than Ta but comparable total new bone area, due to the greater callus area. Therefore, the callus had a considerable contribution for the fixation to become comparable with Ta. Compared with Ta-CaP-ALN, Ta-CaALN was significantly lower in terms of total new bone formation and push-out strength, and it may be explained by its in vitro solubility and release profile. The solubility of CaALN coating is beyond the toxic level of A L N (i.e. l x l O " 4 M) [156]; the free A L N concentration in the buffer during the in vitro release study (i.e. ~9X10"5 M) was lower than the solubility level but close to the toxic level. So the activity of osteoblasts might have been suppressed together with osteoclasts. 130 CD o 3 cp =5 CL 600 500 h 400 300 200 Week 8 R =0.33 A Ingrowth area o Total new bone area _t_ 1 2 3 4 5 _2x Bone area (mm ) (b) Figure 6.12 Correlation between implant push-out force and bone area at (a) week 4 and (b) week 8; results from all implant groups are included. 13! Week 8 For all implant groups, the mean push-out force and strength of 8 week implants were higher than 4 week implants; however, the increases. were achieved in different ways for different groups (Fig 6.8-6.10). Ta and Ta-CaALN showed increase in all of the histomorphometric parameters whereas Ta-CaP-ALN showed decreases in most of them. The different ways of bone mass change can be understood in terms of bone remodeling following Wolffs law [157]. As a general rule for bone fracture healing, woven bone was first laid down as the structural support and was remodeled into more compact and stronger bone when the initial structural stability was achieved. The process is known to be accompanied by an initial increase and a subsequent decrease in callus size. The bone mass (i.e. area) changes observed in this study may be explained as follows. Four weeks after implantation, Ta-CaP-ALN implants achieved good fixation; the ingrowth bone in the superficial pores took an increasing part in load transfer, therefore it became denser and more adapted to the load. On the other hand, the denser bones combined with the stiffness of Ta made the bone islands in the inner pores increasingly stress-shielded and therefore resorbed. The overall effect was the 5% decrease in ingrowth percentage from week 4 to week 8. Similarly, with the load returning to the original tibia trajectory (taken by the ingrowth bone), the contribution of periosteal and endosteal callus to load transfer decreased; their sizes reduced. However, for Ta and Ta-CaALN implants, the fixations at week 4 were probably not sufficient. As the response, bony tissue continued to net-form to stabilize the implant, resulting in larger ingrowth bone and callus area. For Ta and Ta-rn CaALN, the remodeling (i.e. callus area decrease) may occur in a later stage after a sufficient implant fixation is achieved. Therefore, the healing process of Ta-CaP-ALN implants was accelerated over Ta and Ta-CaALN. It may be mentioned that a similar phenomenon was reported in a recent study on porous Ta cylinders implanted in goat femur diaphyses [158]. In that study, the bare Ta showed increasing bone ingrowth up to 24 weeks, whereas the biomimetic carbonate apatite coated Ta showed increasing bone ingrowth from 6 to 12 weeks but decrease from 12 to 24 weeks. Additionally, the ingrowth into the carbonate apatite coated Ta was higher than bare Ta at 6 and 12 weeks, but was surpassed by bare Ta at 24 weeks. The decrease in bone area (from week 4 to week 8) of the Ta-CaP-ALN group showed that the CaP-ALN coating did not effectively inhibit the bone resorption after week 4. This was likely because A L N release was completed with coating erosion. Additionally, A L N should have shorter half-life (and anti-osteoclastic effect) in the new bony tissue than in normal bone, because of its higher driving force of being remodeled [149]. This animal study had some limitations. First, although the implantation in cortical bone allowed reliable calibration of push-out force into push-out strength due to its well-defined thickness, it is different from the clinical situation of joint replacements, where artificial joints are typically implanted in trabecular bones. Second, the use of porous implants, although clinically relevant, made it extremely difficult to evaluate the interfacial stiffness of the bone-implant system. The interfacial stiffness may be an important parameter for characterizing the bone-implant fixation, particularly for the early healing 133 stage. Third, this study only measured the new bone areas, but the pattern of the new bone formation was not analyzed. Two mechanisms of implant-associated new bone formation have been identified in the literature. New bone formation may be either initiated from the surface of the existing host bone and approach the implant (distal osteogenesis) or directly laid down on the implant surface (contact osteogenesis) [159]. The effects of locally delivered bisphosphonates on these factors remain to be evaluated in further studies. The ultimate goal of this study is to develop an ideal bisphosphonates-containing coating that enhances early fixation and delay future osteolysis. Based on the present results, it is expected that the ideal coating should release small dose of drug at the early stage and release larger dose after the immediate bone healing stage. A possible coating may have a more sophisticated structure to combine the advantages of the two coatings. This would be an interesting study in the future. 6. 5 Conclusions This study evaluated the push-out strength and new bone formation of different porous Ta implants in a femoral diaphysis cortical bone model. Four weeks after implantation, the Ta-CaP-ALN implants showed higher push-out force, push-out strength and new bone formation than the control groups. Eight weeks after implantation, the differences among implants were not significant, although the Ta-CaP-ALN implants still displayed the highest push-out strength. From week 4 to week 8, Ta-CaP-ALN implants showed decrease in total bone area, whereas the other implants showed increase in the area; this different trend indicated that the CaP-ALN coating accelerated the implant fixation and 134 bone healing. At both week 4 and 8, Ta-CaALN implants showed similar push-out strength and total new bone area with Ta implants, indicating safety of the CaALN coating. Because the CaALN coating provides the unique possibility of delivering significantly higher drug dose, it warrants further studies to fully exploit this advantage. 135 Chapter 7 Conclusions and Recommendations 7.1 Conclusions This thesis studied the processing and properties of two bisphosphonate-containing coatings on Ti and porous Ta implants, and evaluated their in vivo performances in an animal model. Based on the experimental studies, the following conclusions can be made: 1. A reproducible ELD process was developed to prepare microporous CaP coatings on Ti and Ta, with pore sizes between 0.5-1 um. The coating solutions and deposition potentials were experimentally determined. Coating solution 3 (5.25 mM Ca, 10.5 mM P, pH 5.3) was found to be the optimum solution within this study. ELD in solution 3 resulted in microporous coatings without DCPD and the coating deposition was relatively rapid. XRD and FTIR indicated the coating contained OCP but the possible inclusion of apatitic phase could not be excluded. With deposition time, the coating changed from a cellular morphology to an extended flake morphology; the crystal size and Ca/P ratio increased, and the crystalline disorder seemed to have decreased. ELD current showed a rapid initial decrease followed by a long nearly-constant stage, following the Cottrel equation. Reduction of proton was the main electrode reaction at the first ELD stage, and electrolysis of molecular water became the main reaction with deposition time. Oxygen or nitrate was not critical for the ELD process. Alendronate was chemically adsorbed on the CaP coated Ti and Ta. The in 136 vitro release of alendronate from porous Ta implants was slow, with only 10% eluted after 7 days. 2. The ELD technique was successfully extended to bisphosphonate drugs. Continuous coatings of calcium etidronate and alendronate were deposited at room temperature on flat Ti substrates and porous Ta implants. The chemical structures of the bisphosphonates were not changed by the electrolytic process. The solubilities of the coatings in the buffer solution were 6 x l O " 5 M for Ca-etidronate and 2.5 x l O " 4 M for Ca-alendronate. In vitro study showed drug release from the Ca-alendronate coated porous Ta was completed within 3 days and the alendronate concentration was below the solubility limit. The coatings could provide local release of high dose of bisphosphonate drugs from metallic implants. 3. Four types of porous Ta implants were implanted into the tibial diaphyses of rabbits for 4 weeks and 8 weeks, i.e. the native Ta, Ta ELD coated with CaP (Ta-CaP), Ta ELD coated with CaP and chemically adsorbed with alendronate (Ta-CaP-ALN) and Ta ELD coated with Ca-alendronate coating (Ca-CaALN). After implantation for 4 weeks, the Ta-CaP-ALN implants displayed significantly higher total new bone area and the push-out strength. The Ta-CaALN coating showed generally similar implant fixation and new bone formation to the native Ta implants. After implantation for 8 weeks, although the Ta-CaP-ALN still showed the highest value, there is no significant difference in push-out strength among different implants. From week 4 to week 8, the Ta-CaP-ALN implants showed decrease in total new bone area whereas other implants showed increase in the area. The results therefore indicated the CaP-137 A L N coating accelerated the implant fixation and local bone healing. The Ca-ALN coating did not show significant advantages, nor disadvantages, indicating its safety; but because it provided the unique possibility of delivering dramatically higher doses, further studies are warranted to fully exploit this property. 7.2 Recommendations The following may be recommended for future studies: 1. The bonding strength of the ELD CaP coating to substrates was not studied in this thesis. Although this may not be a significant concern for porous implants, a strong interfacial bonding may be an important requirement for the potential application of the ELD coatings on the surface of non-porous implants. In the literature, the bonding strength of CaP coatings was often evaluated by tensile or shear test on two cylindrical samples joined by glues. However, since the ELD CaP coating was porous, glues may infiltrate into the pores and even directly bond to the substrate and result in artifacts. A recently developed substrate straining test technique may be used to evaluate the bonding strength [160]. 2. The in vivo study in this thesis was conducted in the cortical bone. The well-defined thickness of cortical allowed calibration of the tested push-out force into push-out strength. In addition, porous Ta has potential applications in the repair of cortical bone repair. However, in joint replacements, implants face trabecular bone. To be more clinically relevant, further studies may evaluate the effect of bisphosphonate-138 containing coatings on new bone formation and implant fixation with trabecular bone. In addition, the effect of bisphosphonate on the effect of pattern of new bone formation, namely distal osteogenesis versus contact osteogenesis, deserves to be studied. The dose of bisphosphonate drug chemically adsorbed on the CaP coating obviously depends on the surface area of the coating. With a given surface area, although the adsorbed dose may be increased by increasing the drug concentration of the adsorbing media, it suffers two limitations: (a) the dose would still saturate at fairly low concentration (due to the strong chelation of bisphosphonate with Ca), as indicated by a recent study [140], (b) a higher dose obtained by equilibrating at higher drug concentrations may in turn result in a local in vivo drug concentration exceeding the safety range for osteoblasts and impairing new bone formation, as evidenced in a recent in vivo study [11]. In order to chemically adsorb a high dose but without these disadvantages, two possibilities may be explored. First, the surface area of the CaP coating may be significantly increased by development of a hierarchical topography. This may be realized by ELD deposition of a layer of DCPD with crystal sizes of tens of micrometers, followed by deposition of an apatite/OCP microporous layer. The resulting composite coating may be converted into pure HA by a simple alkaline treatment. The coating prepared by this dual-ELD procedure may have much larger surface area, enabling adsorption of a higher dose of bisphosphonate at low equilibrating drug concentrations. 139 Second, bisphosphonate molecules may be co-depoisted in the 3D CaP matrix, instead of adsorption on the 2D surface. This may be realized by introducing bisphosphonates into the ELD solution at appropriate concentrations. However, since bisphosphonate effectively block CaP crystal nucleation and growth, the concentrations of Ca, phosphate and bisphosphonate should be carefully selected to enable the formation of such composite coatings. The selection may be facilitated by establishing the Ca-P precipitation boundaries by titration in the presence of bisphosphonates. 4. Currently, the Ca-bisphosphonate coating is the only method to locally deliver high dose of bisphosphonate, although it did not show any early-stage advantages in this thesis. An ideal bisphosphonate-containing coating may possess an advanced structure to combine the advantage of the CaP-ALN coating on early fixation and the potential benefit of the Ca-bisphosphonate coating for inhibiting future osteolysis (due to the high dose). This may be realized by encapsulating the Ca-bisphosphonate coating with a biocompatible polymer or inorganic coatings. A CaP coating may be biomimetically prepared on the top of the over-coat for adsorption of bisphosphonate. 5. Some patients may have received bisphosphonate therapy (e.g. for osteoporosis) before being considered for joint replacements. Clinically, some patients may develop resistance to certain bisphosphonates that they have been treated with. The cause of this resistance is unknown, but it can sometimes be overcome by switching to another 140 bisphosphonate (often more potent one, i f possible). Therefore, it would be valuable to ascertain whether a patient is resistance to a local bisphosphonate delivery coating and/or whether this resistance could be overcome by a coating containing another bisphosphonate. A technique may be developed and uses an animal model to answer this question. More valuably, a clinical assay may be developed to screen the resistance for a specific patient. 141 REFERENCES 1. Teeny SM, York SC, Mesko JW, Rea RE, Long-term follow-up care recommendations after total hip and knee arthroplasty: results of the American association of hip and knee surgeons' member survey, J A r t h r o p l a s t y , 2003, 18, 8, 954-962. 2. Pilliar R M , Cementless fixation - toward improved reliability, Orthop Clin North Am, 2005,36, 113-119. 3. Geesink RG, De Groot K, Klein CP, Chemical implant fixation using hydroxyl-apatite coatings. The development of a human total hip prosthesis for chemical fixation to bone using hydroxyl-apatite coatings on titanium substrates, C l i n O r t h o p R e l a t R e s , 1987, 225, 147-170. 4. Geesink RGT, Hydroxyapatite-coated hip implants: experimental studies, in Geesink RGT , Manley MT ed. H y d r o x y a p a t i t e c o a t i n g s in o r t h o p a e d i c s u r g e r y , p 151-170, Raven, New York, 1993. 5. Shanbhag AS, Hasselman CT, Rubash HE, Inhibition of wear debris mediated osteolysis in a canine total hip arthroplasty, C l i n O r t h o p R e l a t R e s , 1997, 344, 33-43. 6. Fleisch H, B i s p h o s p h o n a t e s in b o n e d i s e a s e , Academic Press, San Diego, 4th ed., 2000. 7. Shanbhag AS, Use of bisphosphonates to improve the durability of total joint replacements, J Am Acad Orthop Surg, 2006, 14(4), 215-225. 8. Yoshinari M , Oda Y, Inoue T, Matsuzaka K, Shimono M , Bone response to calcium phosphate-coated and bisphosphonate-immobilized titanium implants, B i o m a t e r i a l s , 2002, 23(14),2879-2885. 142 9. Kajiwara H, Yamaza T, Yoshinari M , Goto T, Iyama S, Atsuta I, Kido M A , Tanaka T, The bisphosphonate pamidronate on the surface of titanium stimulates bone formation around tibial implants in rats, Biomaterials, 2005, 26(6), 581-587. 10. Tengvall P, Skoglund B, Askendal A, Aspenberg P, Surface immobilized bisphosphonate improves stainless-steel screw fixation in rats, Biomaterials, 2004, 25(11), 2133-2138. 11. Peter B, Pioletti DP, Laib S, Bujoli B, Pilet P, Janvier P, Guicheux J, Zambelli PY, Bouler JM, Gauthier O, Calcium phosphate drug delivery system: influence of local zoledronate release on bone implant osteointegration, Bone, 2005, 36(1), 52-60. 12. Tanzer M , Karabasz D, Krygier JJ, Cohen R, Bobyn JD, Bone augmentation around and within porous implants by local bisphosphonate elution, Clin Orthop Relat Res, 2005,441,30-39. 13. Nicholson JW, The chemistry of medical and dental materials, p4, Royal Society of Chemistry, London, 2002. 14. Health Canada, Arthritis in Canada, 2003. 15. Osteoporosis Canada, About osteoporosis, 2007. 16. Papadimitropoulos EA, Coyte PC, Josse RG, Greenwood CE, Current and projected rates of hip fracture in Canada, Can Med Assoc J, 1997, 157(10), 1357-1363. 17. Health Canada, Economic burden of illness in Canada, 1998. 18. American Academy of Orthopedic Surgeons, Osteoarthritis of the hip: A compendium of evidence-based information and resources, 2003. 19. Canadian Institute for Health Information, Hip and knee replacements in Canada -2006 Report. 20. Charnley J, Arthroplasty of the hip - a new operation, Lancet, 1961, 277(7187), 1129-1132. 143 21. Windier M , Klabunde R, Titanium for hip and knee prostheses, in Brunette D M , Tengvall P, Textor M , Thomsen P ed. Titanium in medicine, p703-746, Springer, Berlin, 2001. 22. Schenk R, The corrosion properties of titanium and titanium alloys, in Brunette D M , Tengvall P, Textr M , Thomsen P ed. Titanium in medicine, p 145-170, Springer, Berlin, 2001. 23. Katz JL, Ambrose CG, McMillin C, Spencer P, Orthopedic biomaterials, in Bowllin GL, Wnek G ed. Encyclopedia of biomaterials and biomedical engineering, p i 160-1170, Marcel Dekker, NY, 2006. 24. Branemark PI, Hansson BO, Adell R, Breine U , Lindstrom J, Hallen O, Ohman A, Osseointegrated implants in the treatment of the edentulous jaw: experience from a 10-year period, Scand JPlast Reconstr Surg, 1977, 16, 1-132. 25. Cross MJ, Parish EN, Machan S, Knee joint replacement, in Gowlin GL, Wnek G ed. Encyclopedia of biomaterials and biomedical engineering, p 874-882, Marcel-Dekker, New York, 2004. 26. Petty W, Fixation methods, in Petty W ed., Total joint replacement, p61-74, WB Saunders, Philadelphia, 1991. 27. Gruen TA, McNeice G M , Amstutz HC, Modes of failure of cemented stem-type femoral components: a radiographic analysis of loosening, Clin Orthop Relat Res, 1979, 141, 17-27. 28. Beckenbaugh RD, Ilstrup D M , Total hip arthroplasty, J Bone Joint Surg Am, 1978, 60,3,306-313. 29. Cotterill P, Hunter GA, Tile M , A radiographic analysis of 166 Charnley-Muller total hip arthroplasties, Clin Orthop Relat Res, 1982, 163, 120-126. 30. Cameron HU, Macnab I, Pilliar R M , A porous metal system for joint replacement surgery, Int J Artificial Organs, 1978, 1(2), 104-109. 144 31. Bobyn JD, Ph.D. thesis, University of Toronto, 1980. 32. Furlong R, A new concept of prosthetic fixation, in Geesink RGT, Manley MT ed., Hydroxyapatite coatings in orthopaedic surgery, p 25-31, Raven Press, New York, 1993. 33. Pilliar R M , Porous surfaced metallic implants for orthopedic application, J Biomed Mater Res, 1987, 21(A1 Suppl), 1-33. 34. Bobyn JD, Stackpool GJ, Hacking SA, Tanzer M , Krygier JJ, Characteristics of bone ingrowth and interface mechanics of a new porous tantalum biomaterial, J Bone Joint Surg Br, 1999, 81(5), 907-914. 35. Medlin DJ, Scrafton J, Shetty R, Metallurgical attachment of a porous tantalum foam to a titanium substrate for orthopedic applications, J ASTM Intl, 2005, 2(10), 1-10. 36. Zardiackas LD, Parsell DE, Dillon LD, Mitchell DW, Nunnery L A , Poggie R, Structure, metallurgy, and mechanical properties of a porous tantalum foam, J Biomed Mater Res, 2001, 58(2), 180-187. 37. Bobyn JD, Poggie RA, Krygier JJ, Lewallen DG, Hanssen AD, Lewis RJ, Unger AS, O'keefe TJ, Christie MJ, Nasser S, Wood JE, Stullberg SD, Tanzer M , Clinical validation of a structural porous tantalum biomaterial for adult reconstruction, J Bone Joint Surg Am, 2004, 86A(Suppl 2), 123-129. 38. Jasty M , Bragdon C, Burke D, O'connor D, Lowenstein J, Harris WH, In vivo skeletal responses to porous-surfaced implants subjected to small induced motions, J Bone Joint Surg Am, 1997,79(5), 707-714. 39. Soballe K, Hansen ES, Rasmussen HB, Biinger C, The effects of osteoporosis, bone deficiency, bone grafting, and micromotion on fixation of porous-coated versus hydroxyapatite-coated implants, in Geesink RGT, Manley MT ed. Hydroxyapatite coating in orthopedic surgery, p 107-136, Raven, New York, 1993. 145 40. Linder L, An overview of the histology of the principal interface types in orthopedic surgery today, in Older J ed., Implant Bone Interface, Springer-Verlag, London, 1990. 41. Freeman M A , Plante-Bordeneuve P, Early migration and late aseptic failure of proximal femoral prostheses, J Bone Joint Surg Br, 1994, 76(3), 432-438. 42. Ryd L, Albrektsson BEJ, Caisson L, Dansgard F, Herberts P, Lindstrand A, Regener L and Toksvig-Larsen S, Roentgen stereophotogrammetric analysis as a predictor of mechanical of knee prostheses, J Bone Joint Surg Br, 1995, 77(3), 377-383. 43. Hofmann A A , Bachus K N , Bloebaum RD, Comparative study of human cancellous bone remodeling to titanium and hydroxyapatite coated implants, J Arthroplasty, 1993, 8(2), 157-166. 44. Swedish National Hip Arthroplasty Register, Annual Report 2005. 45. Im G, Sethi RK, Rubash HE, Shanbhag AS, Osteolysis in total hip arthroplasty: biological and clinical aspects, in Sinha R K ed. Hip replacement current trends and controversies, p 177-212, Marcel-Dekker, New York, 2002. 46. Harris WH, The problem is osteolysis, Clin Orthop Relat Res, 1995, 311, 46-53. 47. Harris WH, Wear and periprosthetic osteolysis: the problem, Clin Orthop Relat Res, 2001,393,66-70. 48. Margevicius KJ , Bauer TW, McMahon JT, Brown SA, Meritt K, Isolation and characterization of debris in membranes around total joint prostheses, J Bone Joint Surg Am, 1994, 76(11), 1664-1675. 49. Brown WE, Chow LC, Chemical properties of bone mineral, Ann Rev Mater Sci, 1976, 6,213-236. 50. Jarcho M , Kay JF, Gumaer K, Doremus RH, Drobeck HP, Tissue, cellular and subcellular events at a bone-ceramic hydroxyapatite interface, J Bioeng, 1977, 1, 2, 79-92. 146 51. Tracy B, Doremus R, Direct electron microscopy studies of the bone hydroxyapatite interface, J Biomed Mater Res, 1984, 18, 7, 719-726. 52. Manley MT, Calcium phosphate biomaterials: a review of the literature, in Geesink RGT,Manley MT ed. Hydroxyapatite coatings in orthopaedic surgery, p 1-24, Raven, New York, 1993. 53. Porter AE, Patel N , Best S, Hydroxyapatite, in Bowllin GL, Wnek G ed., Encyclopedia of biomaterials and biomedical engineering, p 1-12, Marcel-Dekker, NY, 2006. 54. De Groot K, Geesink R, Klein CP, Serekian P, Plasma-sprayed coatings of hydroxyapatite, J Biomed Mater Res, 1987,21(12), 1375-1381. 55. Tsui Y C , Doyle C, Clyne TW, Plasma sprayed hydroxyapatite coatings on titanium substrates part 1: Mechanical properties and residual stress levels, Biomaterials, 1998, 19(22), 2015-2029. 56. Tsui Y C , Doyle C, Clyne TW, Plasma sprayed hydroxyapatite coatings on titanium substrates part 2: optimisation of coating properties, Biomateirals, 1998, 19(22), 2031-2043. 57. Gruner H, Thermal spray coatings on titanium, in Brunette D M , Tengvall P, Textor M , Thomsen P ed. Titanium in medicine, p 375-416, Springer, Berlin, 2001. 58. Stewart M , Welter JF, Goldberg V M , Effect of hydroxyapatite/tricalcium-phosphate coating on osseointegration of plasma-sprayed titanium alloy implants, J Biomed Mater Res A, 2004, 69(1), 1-10. 59. Gross K A , Berndt CC, Thermal processing of hydroxyapatite for coating production, J Biomed Mater Res, 1998, 39(4), 580-587. 60. Van Blitterswijk CA, Bovell YP, Flach JS, Leenders H, van den Brink I, de Bruijn J, Variations in hydroxyapatite crystallinity: effects on interface reactions, in Geesink 147 RGT, Manley MT ed. Hydroxyapatite coatings in orthopaedic surgery, p 33-47, Raven Press, New York, 1993. • 61. Paschalis EP, Zhao Q, Tucker BE, Mukhopadhayay S, Bearcroft JA, Beals NB, Spector M , Nancollas GH, Degradation potential of plasma-sprayed hydroxyapatite coated titanium implants, JBiomed Mater Res, 1995, 29(12), 1499-1505. 62. Brossa F, Cigada A, Chiesa R, Paracchini L, Consonni C, Post-deposition treatment effects on hydroxyapatite vacuum plasma spray coatings, J Mater Sci Mater Med, 1994, 5(12), 855-857. 63. Jaffe WL, Scott DF, Current concepts review - total hip arthroplasty with hydroxyapatite-coated prostheses, J Bone Joint Surg Am, 1996, 78(12), 1918-1934. 64. Dumbleton J, Manley MT, Hydroxyapatite-coated prostheses in total hip and knee arthroplasty, J Bone Joint Surg Am, 2004, 86(11), 2526-2540. 65. Havelin LI, Engesaeter B, Espehaug B, Furnes O, Lie SA, Vollset SE, The norwegian arthroplasty register: 11 years and 73,000 arthroplasties, Acta Orthop Scand, 2000, 71(4), 337-353. 66. Lacefield WR, Hydroxyapatite coatings, in Ducheyne P, Lemons JE ed. Bioceramics: materials characteristics versus in vivo behavior, Ann New York Academy Sci, 1988, 523, 72-80. 67. Ji HX, Ponton CB, Marquis PM, Microstructural characterization of Hydroxyapatite coating on titanium, J Mater Sci Mater Med, 1992, 3, 283-287. 68. Yang Y C , Chang E, Influence of residual stress on bonding strength and fracture of plasma-sprayed hydroxyapatite coatings on Ti6A14V substrate, Biomaterials, 2001, 22(13), 1827-1836. 69. Tonino A, Oosterbos C, Rahmy A, Therin M , Doyle C, Hydroxyapatite-coated acetabular components: histological and histomorphometric analysis of six cups 148 retrieved at autopsy between three and seven years after successful implantation, J Bone Joint Surg Am, 2001, 83A(6), 817-825. 70. Gross K A , Ray N , The contribution of coating microstructure to degradation and particle release in hydroxyapatite coated prostheses, Rokkum M , J Biomed Mater Res, 2002, 63(2), 106-114. 71. Mulllin JW, Crystallization, pi81-195, Butterworth-Heinemann, Boston, 4th edn, 2001. 72. Bunker BC, Rieke PC, Tarasevich BJ, Campbell AA, Fryxell GE, Graff GL, Song L, Liu J, Virden JW, McVay GL, Ceramic thin-film formation on functionalized interfaces through biomimetic processing, Science, 1994, 264(5155), 48-55. 73. Mann S, Biomineralization: Principles and Concepts in Bioinorganic Chemistry, ed. Mann S, p 6-122, Oxford University Press, Oxford, 2001. 74. Paks GA, The isoelectric points of solid oxides, solid hydroxides and aqueous hydroxo complex systems, Chem Review, 1965, 65(2), 177-198. 75. Campbell A A , Fryxell GE, Linehan GC, Graff GL, Surface induced mineralization: a new method of producing calcium phosphate coatings, J Biomed Mater Res, 1996, 32(1), 111-118. 76. Campbell A A , Song L, L i XS, Nelson BJ, Bottoni C, Brooks DE, DeJong ES, Development, characterization, and anti-microbial efficacy of hydroxyapatite-chlorhexidine coatings produced by surface-induced mineralization. J Biomed Mater Res B, 2000, 53(4), 400-407. 77. Liu Q, Ding J, Mante FK, Wunder SL, Baran GR, The role of surface fundtional groups in calcium phosphate nucleation on titanium foil: a self-assembled monolayer technique, Biomaterials, 2002, 23(5), 3103-3111. 78. Elliott JC, Structure and chemistry of the apatites and other calcium orthophosphates, p 111-186, p 191-275, Elsevier Science, Amsterdam, 1994. 149 79. Soten I, Ozin GA, Porous Hydroxyapatite-dodecylphosphate composite film on titania-titanium surface, J Mater Chem, 1999, 9, 3, 703-710. 80. Kim H M , Miyaji F, Kokubo T, Preparation of bioactive Ti and its alloys via simple chemical treatment, JBiomed Mater Res, 1996, 32(3), 409-417. 81. Kim H M , Miyaji F, Kokubo T, Graded surface structure of bioactive titanium prepared by chemical treatment, J Biomed Mater Res, 1999, 45(2), 100-107 82. Miyazaki T, Kim H M , Miyaji F, Kokubo T, Kato H, Nakamura T, Bioactive tantalum metal prepared by NaOH treatment, J Biomed Mater Res, 2000, 50(1), 35-42. 83. Ohtsuki C, Iida H, Hayakawa S, Osaka A, Bioactivity of titanium treated with hydrogen peroxide solutions containing metal chlorides, J Biomed Mater Res, 1997, 35(1), 39-74. 84. Wen HB, Liu Q, De Wijn JR, De Groot K, Cu FZ, Preparation of bioactive microporous titanium surface by a new two-step chemical treatment, J Mater Sci Mater Med, 1998, 9(3), 121-128. 85. Yan WQ, Nakamura T, Kobayashi M , Kim H M , Miyaji F, Kokubo T, Bonding of chemically treated titanium implants to bone, J Biomed Mater Res, 1997, 37(2), 267-275. 86. Zitelli JP, Higham P, A novel method for solution deposition of hydroxyapatite onto three dimentionally porous metallic surfaces, in Materials Research Symposium (Mineralization in natural and synthetic biomaterials), Materials Research Society, Warrendale, 1999, (599), 117-128. 87. Barrere F, Snel M M , Van Blitterswijk CA, De Groot K, Layrolle P, Nano-scale study of the nucleation and growth of calcium phosphate coating on titanium implants, Bomaterials, 2004, 25(14), 2901-2910. 150 88. Barrere F, Van Blitterswijk CA, De Groot K, Layrolle P, Nucleation of biomimetic Ca-P coatings on Ti6A14V from a SBF><5 solution: influence of magnesium, Bomaterials, 2002, 23(10), 2211-2220. 89. L i P, Biomimetic nano-apatite coating capable of promoting bone ingrowth, J Biomed Mater Res A, 2003, 66(1), 79-85. 90. Zhitomirsky I, Cathodic electrodeposition of ceramic and organoceramic materials fundamental aspects, Adv Colloid Interface Sci, 2002, 97(1-3), 277-315. 91. Redepenning J, Electrocrystallization of brushite coatings on prosthetic alloys, Chem Mater, 1990, 2(6), 625-627. 92. Therese GHA, Kamath PV, Subbana GN, Novel synthetic routes to calcium phosphate coatings, J Mater Chem, 1998, 8, 405-408. 93. Shirkhanzadeh M , Direct formation of nanophase hydroxyapatite on cathodically polarized electrodes, J Mater Sci Mater Med, 1998, 9(2), 67-72. 94. RoBler S, Sewing A, Stolzel M , Born R, Scharnwebber D, Dard M , Worch H, Electrochemically assisted deposition of thin calcium phosphate coatings at near-physiological pH and temperature, J Biomed Mater Res A, 2002, 64(4), 655-663. 95. Peaker AJS, Czernuszka J, The effect of electric field on the formation of hydroxyapatite coatings, Thin Solid Films, 1996, 287(1-2), 174-183 96. Yen SK, Lin SM, Cathodic reactions of electrolytic hydroxyapatite coating on pure titanium, Mater Chem Phys, 2002, 77(1), 70-76. 97. Baudler M , Schellenberg D, Beitrage zur Chemie des Phosphors X X I V Elektrolytische Untersuchungen von Phosphorsauren in waBriger Losung, Z Anorg Allg Chem, 1965, 340(3-4), 113-125. 98. Kumar M , Dasarathy H, Riley C, Electrodeposition of brushite coatings and their transformation to hydroxyapatite in aqueous solutions, J Biomed Mater Res, 1999, 45(4), 302-310. 151 99. US Patent 5723038. 100. Costa L, Mechanisms of crosslinking and oxidative degradation of UHMWPE, in Kurtz SM ed., The UHMWPW handbook, p 245-262, Elsevier Academic Press, San Diego, 2004. 101. McKellop H, Shen FW, Lu B, Campbell P, Salovery R, Development of an extremely wear-resistant ultra high molecular weight polyethylene for total hip replacements, J Orthop Res, 1999, 17(2), 157-167. 102. Muratoglu OK, Bragdon CR, O'Connor DO, Jasty MJ, Harris WH, A novel method of cross-linking ultrahigh molecular weight polyethylene to improve wear, reduce oxidation and retain mechanical properties, J Arthroplasty, 2001, 16(2), 149-160. 103. Muratoglu KO, Kurtz SM, Alternative bearing surface in hip replacement, in Sinha R K ed., Hip replacement current trends and controversies, p 28-35, Marcel-Dekker, New York, 2002. 104. Riecker CB, Kottig P, Schon R, Windier M , Syss U, Clinical tribological performance of 144 metal-on-metal hip articulations, in Rieker CB, Windier M , Wyss U ed. Metasul: A metal-on-metal bearing, p83-92, Huber & Lang, Bern, 1999. 105. Tipper JL, Firkins PL, Ingham E, Fisher J, Stone M H , Farrar R, Quantitative analysis of the wear and wear debris from low and high carbon content cobalt chrome alloys used in metal on metal total hip replacement, J Mater Sci Mater Med, 1999, 10(6), 353-362. 106. Lhotka C, Szekeres T, Steffan I, Zhuber K, Zweymuller K, Four-year study of cobalt and chromium blood levels in patients managed with two different metal-on-metal total hip replacements, J Orthop Res, 2003, 21(2), 189-195. 107. Jacobs JJ, Skipor A K , Patterson L M , Metal release in patients who have had a primary total hip arthroplasty: a prospective, controlled, longitudinal study, J Bone Joint Surg Am, 1998, 80(10), 1447-1558. 152 108. Raisz LG, Physiology and pathophysiology of bone remodeling, Clin Chem, 1999, 45(8), 1353-1358. 109. Khan SA, Vasikaran S, Kline WF, Elimination and biochemical response to intravenous alendronate in postmenopausal osteoporosis, J Bone Mineral Res, 1997, 12(10), 1700-1707. 110. Eekhoff 'ME, Zwindermann A H , Haverkort D M , Cremers SC, Hamdy NA, Papapoulos SE, Determinants of induction and duration of remission of Paget's disease of bone after bisphosphonate therapy, Bone, 2003, 33(5), 831-838. 111. Vasikaran SD, Bisphosphonates: an overview with special reference to alendronate, Ann Clin Biochem, 2001, 38(6), 608-623. 112. Zou X , Xue Q, L i H, Bunger M , Lind M , Bunge C, Effect of alendronate on bone ingrowth into porous tantalum and carbon fiber interbody devices: an experimental study on spinal fusion in pigs, Acta Orthop Scand, 2003, 74(5), 596-603. 113. Lin JH, Bisphosphonates: a review of their pharmacokinetic properties, Bone, 1996, 18(2), 75-85. 114. Bragdon CR, Doherty A M , Jasty M , Rubash M , Harris WH, Effect of oral alendronate on net ingrowth into canine cementless total hips, J Arthroplasty, 2005, 20(2), 258-263. 115. Bilezikian JP, Osteonecrosis of the jaw - do bisphosphonates pose a risk, N Engl J Med, 2006, 355(22), 2278-2281. 116. Denissen H, Montanari C, Martinetti R, Van Lingen A, Van den Hooff A, Alveolar bone response to submerged bisphosphonate complexed hydroxyapatite implants, J Periodontal, 2000, 71(2), 279-286. 117. Meraw SJ, Reeve C M , Wollan PC, Use of alendronate in peri-implant defect regeneration, J Periodontal, 1999, 70(2), 151-158. 153 118. Meraw SJ, Reeve C M , Qualitative analysis of peripheral peri-implant bone and influence of alendronate sodium on early bone regeneration, J Periodontal, 1999, 70(10),1228-1233. 119. International Standard Organization, Biological evaluation of medical devices - Part 6: Tests for local effects after implantation. 120. Hermanson GT, Bioconjugate techniques, p 137-138, 298-312, Academic Press, San Diego, 1996. 121. Johnston TP, Schoen FJ, Levy RJ, Prevention of calcification of bioprosthetic heart valve leaflets by Ca 2 + diphosphonate pre-treatment, J Pharm Sci, 1988, 77(9), 740-744. 122. Fleisch G, Hauffe SA, Determination of the bisphosphonate pamidronate disodium in urine by pre-column derivatization with fluorescamine, high-performance liquid chromatography and fluorescence detection, J Chromatogr-Biomed App, 1989, 489(2), 446-451. 123. Tovborg JA, Rathlev J, Calcium hydrogen orthophosphate 2-hydrate and calcium hydrogen orthophosphate, Inorg Synthe, 1953, 4, 19-22. 124. Elliott JC, Structure and chemistry of the apatites and other calcium orthophosphates, p 20-21, p 30, pi69-173, Elsevier Science, Amsterdam, 1994. 125. Berry EE, Baddiel CB, The infra-red spectrum of dicalcium phosphate dehydrate, Spectrochim Acta, 23A, 2089-2097. 126. Fowler BO, Moreno EC, Brown WE, Infra-red spectra of hydroxyapatite, octacalcium phosphate and pyrolyzed octacalcium phosphate, Arch Oral Biol, 1966, 11,477-492. 127. Fowler BO, Infrared studies of apatites I: vibrational assignments of calcium, strontium, and barium hydroxyapatite utilizing isotopic substitution, Inorg Chem, 1974, 13(1), 194-207. 154 128. Berry EE, Baddiel CB, Some assignments in the infra-red spectrum of octacalcium phosphate, Spectrochim Acta, 1967, 23A, 1781-1792. 129. Elliott JC, Structure and chemistry of the apatites, and other calcium orthophosphates, p 20, Elsevier, Amsterdam, 1994. 130. Elliott JC, Amorphous calcium phosphates, in Elliott JC ed. Structure and chemistry of the apatites and other calcium orthophosphates, p 58-60, Elsevier Science, Amsterdam, 1994. 131. Matsumoto Y , Morikawa T, Adachi H, Hombo J, A new preparation method of barium titanate perovskite film using electrochemical reduction, Materials Research Bulletin, 1992, 27(11), 1319-1323. 132. Girault HH, Analytical and physical electrochemistry, p 375-380, Marcel-Dekker, New York, 2004. 133. Girault HH., Analytical and physical electrochemistry, p 271-275, Marcel-Dekker, New York, 2004. 134. Girault HH., Analytical and physical electrochemistry, p 301-304, Marcel-Dekker, New York, 2004 135. Montero-Ocampo C, Villegas D, Veleva L, Controlled potential electrodeposition of calcium phosphate on Ti6A14V, JElectrochem Soc, 2005, 152, 10, 692-695. 136. Chen GZ, Fray DJ, Farthing TW, Direct electrochemical reduction of titanium dioxide to titanium in molten calcium chloride, Nature, 2000, 407, 361-364. 137. Platides C, Kyriacou G, Electrochemical reduction of nitrate ion on various cathodes - reaction kinetics on bronze cathode, J Applied Electrochem, 2005, 35(5), 421-427. 138. Lide DR, CRC Handbook of chemistry and physics, 87th edn., CRC Press, Baton Rouge, 2006. 155 139. Zhang JM, Lin CJ, Feng ZD, Tian ZW, Mechanistic studies of electrodeposition for bioceramic coatings of calcium phosphates by an in situ pH microsensor technique, J Electroanalytic Chem, 1998, 452(2), 235-240. 140. Roussiereab H, Montavon G, La'iba S, Janviera P, Alonsod B, Fayond F, Petita M , Massiotd D, Bouler J and Bujoli B, Hybrid materials applied to biotechnologies: coating of calcium phosphates for the design of implants active against bone resorption disorders, J Mater Chem, 2005, 15(35&36), 3869-3875 141. Fleisch H, Bisphosphonates in bone diseases, p 29-33, Academic Press, San Diego, 4th edn, 2000. 142. Den Hartigh J, Langebroek R, Vermeij P, Ion-exchange liquid chromatographic analysis of bisphosphonates in pharmaceutical preparations, J Pharm Biomed Anal, 1993, 11(10), 977-983. 143. Tsai EW, Ip DP, Brooks M A , Determination of alendronate in pharmaceutical dosage formulations by ion chromatography with conductivity detection, J Chromatogr, 1992, 596(2), 217-224. 144. Amerian Society for Testing of Materials, A STM standard 6501-99: standard test method for phosphonate in brines. 145. Thomas LC, Interpretation of the infrared spectra of organo-phosphorous compounds, Heydon, London, 1974. 146. Reinholz, GG, Getz B, Pederson L, Sanders ES, Subramaniam M , Ingle JN, Spelsberg TN, Bisphosphonates directly regulate cell proliferation, differentiation, and gene expression in human osteoblasts, Cancer Res, 2000, 60(21), 6001-6007. 147. Garcia-Moreno C, Serrano S, Nacher M , Farre M , Diez A, Marinoso M L , Carbonell J, Mellibovsky L, Nogues X, Ballester J, Aubia J, Effect of alendronate on cultured normal human osteoblasts, Bone, 1998, 22(3), 233-239. 156 148. Bergstrom JD, Bostedor RG, Masarachia PJ, Reszka AA, Rodan G, Alendronate is a specific, nanomolar inhibitor of farnesyl diphosphonate synthase, Arch Biochem Biophys, 2000, 373(1), 231-241. 149. Schell H, Lienau J, Epari DR, Seebeck P, Exner C, Muchow S, Bragulla H, Haas NP and Duda GN, Osteoclastic activity begins early and increases over the course of bone healing, Bone, 2006, 38(4), 547-554. 150. Aro HT, Wippermann BW, Hodgson SF, Chao EY, Internal remodeling of periosteal new bone during fracture healing, J Orthop Res, 1990, 8(2), 238-246. 151. Peter CP, Cook WO, Nunamaker D M , Provost MT, Seedor JG, Rodan GA, Effect of alendronate on fracture healing and bone remodeling in dogs, J Orthop Res, 1996, 14(1), 74-79. 152. Koivukangas A, Tuukkanen J, Kippo K, Jamsa T, Hannuniemi R, Pasanen I, Vaananen K, Jalovaara P, Long-term administration of clodronate does not prevent fracture healing in rats, Clin Orthop Relat Res, 2003, 408, 268-278. 153. L i J, Mori S, Kaji Y, Kawanishi J, Akiyama T, Norimatsu H, Concentration of bisphosphonate (incardonate) in callus area and its effects on fracture healing in rats, J Bone Miner Res, 2000, 15(10), 2042-2051. 154. Im G, Qureshi SA, Kenney J, Rubash HE, Shanbhag AS, Osteoblast proliferation and maturation by bisphosphonates, Biomaterials, 2004, 25(18), 4105-4115. 155. Hu Y X , M.A.Sc Thesis, The University of British Columbia, 2007; a manuscript submitted for publication. 156. Reinholz GG, Getz B, Pederson L, Sanders ES, Subramaniam M , Ingle JN, Spelsberg TC, Bisphosphonates directly regulate cell proliferation differentiation and gene expression in human osteoblasts, Cancer Res, 2000, 60(21), 6001-6007. 157. Park JB, Lakes RS, Biomaterials: an introduction, p204-206, 2nd edn, Plenum Press, New York, 1992. 157 158. Barrere F, Van Der Valk C M , Meijer G, Dalmeijer RA, De Groot K, Layrolle P, Osteointegration of biomimetic apatite coating applied onto dense and porous metal implants in femurs of goats, J Biomed Mater Res B, 2003, 67(1), 655-665. 159. Davies JE, Understanding peri-implant endosseous healing, J Dent Edu, 2003, 67(8), 932-949. 160. Gan L, Wang J, Pilliar R M , Evaluating interface strength of calcium phosphate sol-gel-derived thin films to Ti6A14V substrate, Biomaterials, 2005, 26(2), 189-96. 158 LIST O F PUBLICATIONS 1. Duan K, Wang R, Surface modification of bone implants through wet chemistry, Journal of Materials Chemistry, 2006, 24: 2309-2321. 2. Duan K, Fan Y, Wang R, Electrolytic deposition of calcium etidronate drug coating on titanium substrate, Journal of Biomedical Materials Research B, 2005, 72: 43-51. 3. Duan K, Fan Y, Wang R, Electrochemical deposition and patterning of calcium phosphate bioceramic coating, Ceramic Transactions, 2003, 147: 53-61. 4. Fan Y, Duan K, Wang R, A composite coating by electrolysis induced collagen self assembly and calcium phosphate mineralization, Biomaterials, 2005, 26: 1623-1632. 5. Duan K, Hu Y, Long K, Toms A, Burt H, Oxland TR, Masri BA, Duncan CP, Garbuz DS, Wang R, The effect of alendronate-containing coatings on early fixation and new bone formation of porous tantalum in a rabbit cortical bone model, to be submitted to Journal of Biomedical Materials Research. 6. Garbuz DS, Hu Y, Kim WY, Duan K, Masri BA, Oxland TR, Burt H, Wang R, DuncanCP, Gap filling and enhanced osteoconduction associated with Alendronate coated porous tantalum, submitted to Journal of Bone and Joint Surgery (American Volume), currently under first revision. 159 APPENDICES Appendix A Some Fundamental Relations between Current and Electrode Potential The current passed at an electrode reflects the rate of electron transfer at the electrode-solution interface. This appendix briefly reviews some fundamental relations between current and electrode potential involved in the experiments of this thesis. A complete, mathematical treatment of all the followings can be found in Ref [132-134]. A . l Electrode-Solution Interface When an electrode is in contact with an electrolyte, electron transfer happens between them and this forms a charged double layer. There are several models for this double layer. The earliest model is Helmhotz model. It simply treats the double layer as a simple parallel capacitor; the charge on the electrode surface is balanced by ions of an opposite sign but equal number of charge (i.e. counterions). The distance between the electrode surface and the counterions is the radius of the counterion. The potential gradient is linear from the electrode surface to the center of the counterion. A modern and widely used model is Guy-Chapman-Stern model. This model considers the counterion as two parts, compact layer and diffuse layer. The compact layer consists of counterions adsorbed to the electrode surface, but the charge of the counterions is not adequate to neutralize the charge on the electrode surface; this layer is also known as Stern layer. The diffuse layer consists of 160 excess counterions (over other ions), and this layer provides the rest of the charge to neutralize the charge on the electrode surface. When the electrode is in contact with the solution, but in an open circuit state, a stable electrode potential is spontaneously established. This potential is known as equilibrium potential. When a potential different from the quilibrium potential is applied on the electrode, the difference between the two potentials (i.e. overpotential) is the driving force for electrode reactions, and currents. A.2 Current Controlled by Electrode Potential — Tafel law The current may be controlled by two factors, the rate of the electrode reaction at the electrode-solution interface, and the rate of mass transport to the electrode surface. If one rate is much lower than the other rate, it controls the current. If an electroactive species is present in a high concentration, or even is the solvent phase itself, the concentration of the species at the electrode surface may be sufficient at any time (non-depletable). Then the current is only controlled by the potential applied to the electrode. The current passed is the result of electron transfer reactions at the electrode-solution interface. According to the principle of chemical equilibrium, the net electrode reaction is further considered as two components: cathodic reaction and anodic reaction. The net current is the net results of the two reactions. Treating each reaction with transition state theory, the rate of each reaction can be described by an exponential law: . i a = k a x e x p ( - ^ ) i c = k c xexp(-^-f) where i is current, k is the rate constant, AG is the activation energy, R is the ideal gas constant, T is temperature (in Kelvin); the subscript a means anodic, and the subscript c means cathodic. Further considering the reaction path, it can be shown that the activation energies of both reactions are related to the overpotential applied to the electrode. The difference between the two activation energies relates to the change of Gibbs free energy of the reaction. Assuming the cathodic reaction and the anodic reaction follow identical but opposite reaction paths, the Gibbs free energy change of them are also identical but with opposite signs. The net current then can be written as the sum of the two reactions: i = JoX[exp(^) ] -exp( |^ ) ] where j 0 , A , B are constants, and r| is the overpotential. This equation is known as Butler-Volmer's equation. The two exponential components represent the two counteracting currents. In reality, usually one of the exponential component is much higher than the other; under this condition the equation can be simplified into a single exponential form: i = a x exp(bn) 162 This is widely known as Tafel's law. It indicates that when the current is controlled by the electrode potential, the current~overpotential follows an exponential relationship. A .3 Current Controlled by Electrode Potential and Diffusion — Linear Potential Sweep Linear potential sweep means linearly scan of the electrode potential from a starting value to an end value. This experiment is called linear sweep voltammetry (LSV), and the current-potential result of this scan is known as linear sweep voltammogram. The mathematical treatment of LSV is lengthy and sophisticated [131]. The final result is: i = nFAC x VrcDa x x(ai) nFv a = RT where i is the current, n is the number of electrons transferred, F is Faraday constant, A is the electrode surface area, C is the concentration of the electroactive species, D is the diffusion coefficient, and % is a specially defined dimensionless function; v is the scanning rate (V/second), R is ideal gas constant and T is the temperature in Kelvin. The shape of a LSV shows a peak. The ascending part of the peak indicates rapid increase of current with potential, corresponding to a potential-controlled current stage. With scanning, the concentration of the electroactive species at the electrode surface gradually decreased, and the current undergoes a transition to a diffusion-controlled stage. 163 The current then decreases at a certain potential value, appearing as a peak. On LSV, the oxidation/reduction of different species shows as separate peaks. LSV is useful in idenfication of reaction mechanisms. A. 4 Current Controlled by Diffusion at A Constant Potential — Cottrel Equation When a potential negative enough is applied to a planar cathode (or, a potential positive enough is applied to an anode), an electroactive species (e.g. H + ) at the electrode surface may be rapidly consumed by electrode reactions, and its concentration is reduced to almost zero. The current then must be sustained, and controlled, by transport of the species from the bulk solution. Again, because the potential is sufficient to deplete the species, all of the species transported from the bulk solution are consumed by the electrode reaction. In addition, if the solution is static, diffusion is the only mechanism of mass transport and it follows Fick's laws. The following differential equation applies: i = n F A D « M dx where C(0, t) is the concentration of the species on the electrode surface and at the time oft. 164 In this model, the following boundary conditions are satisfied: For t=0 and x>0, C=C 0 For t>0 and x=0, C=0 For t>0 and x^oo, C=0 where CO means the bulk concentration of the species. The solution of the differential equation gives the equation: i = nFAC 0 This equation is known as Cottrel equation. It indicates that when a potential ramp of sufficient magnitude is applied to a planar electrode in a static solution, the electrode 1/2 current decays linear with f . 165 A P P E N D I X B - Calibration Curves of Bisphosphonate Assays Figure B.l Calibration curve of fluorescent HPLC for alendronate, with concentration of 0 to 10 u.g/ml (mean ±SD, n=3, error bars too small to be seen). Mobile phase: 97:3 1 mM EDTA (pH 6.5): methanol, 1 ml/min, 3.9X150 mm C18 column, excitation 395 nm, emission 480 nm. 166 Figure B.2 Calibration curve of ion chromatography for alendronate, with concentration of 2.5 to 30 uM (n=l). Eluent: 28 mM NaOH, 1 ml/minute. o o ' • 1 ' 1 ' 1 • 1 ' 1 ' 1 0 1 2 3 4 5 6 Phosphate Concentration (x10~5 M) Figure B.3 Calibration curve of spectrophotometry for alendronate, with concentration of 10 to 50 uM (mean ± SD, n=3). Assay procedures following ASTM 6901-99. Absorbance measured at wavelength: 725 nm. 168 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0078533/manifest

Comment

Related Items