Handbook of Biomineralization

Handbook of Biomineralization Edited by Matthias Epple and Edmund Baüerlein Handbook of Biomineralization. Edited by ...

Author: Edmund Bäuerlein | Peter Behrens | Matthias Epple | Jeremy Pickett-Heaps | Stephen Mann | Wolfgang Pompe

42 downloads 1554 Views 37MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form

DOWNLOAD PDF

Handbook of Biomineralization Edited by Matthias Epple and Edmund Baüerlein

Handbook of Biomineralization. Edited by M. Epple and E. Baüerlein Copyright 8 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ISBN: 978-3-527-31806-3

Handbook of Biomineralization Medical and Clinical Aspects

Edited by Matthias Epple and Edmund Baüerlein

The Editors Prof. Dr. MatthiasËpple University of Duisburg-Essen Inorganic Chemistry Universitätsstr. 2 45141 Essen Germany

Prof. Dr. Edmund Baüerlein Max-Planck-Institute for Biochemistry Department of Membrane Biochemistry Am Klopferspitz 18 A 82152 Planegg Germany ¨ Cover Illustration (designed by Felix Bäuerlein) (Top right, Bottom left and Bottom right designed by Felix Baeuerlein) Top left: Heterodont molar of carnivorous mammals (wolf, no functional wear) with no exposed dentin and very small pulp chamber. (P. Gaengler, W. H. Arnold, Chap. 14. Fig. 14.1c). Top right: Apatite formed on TiO2 gel in simulated body fluid (SBF). (T. Kokubo, H. Takadama. Chap. 7. Fig. 7.4(2)). Bottom left: An implant manufactured by hot pressing and gas flushing for cranial reconstruction with gradients in composition and spatially different porosity. (M. Epple, Chap. 6) Bottom right: Calcified lung metastases of a primary colorectal adenocarcinoma. Light microscopic image, HE-stain, metastases with typical structure of colon (*). typical lung structure is not present, ossification (**). (Inge Schmitz. Chap. 18, Fig. 18.7) Handbook of Biomineralization Biological Aspects and Structure Formation: ISBN 978-3-527-31804-9 Biomimetic and Bioinspired Chemistry: ISBN 978-3-527-31805-6 Medical and Clinical Aspects: ISBN 978-3-527-31806-3 Set (3 volumes): ISBN 978-3-527-31641-0

9 All books published by Wiley-VCH are carefully produced. Nevertheless, authors, editors, and publisher do not warrant the information contained in these books, including this book, to be free of errors. Readers are advised to keep in mind that statements, data, illustrations, procedural details or other items may inadvertently be inaccurate. Library of Congress Card No.: applied for British Library Cataloguing-in-Publication Data A catalogue record for this book is available from the British Library. Bibliographic information published by the Deutsche Nationalbibliothek Die Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data are available in the Internet at hhttp://dnb.d-nb.dei. 8 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim All rights reserved (including those of translation into other languages). No part of this book may be reproduced in any form – by photoprinting, microfilm, or any other means – nor transmitted or translated into a machine language without written permission from the publishers. Registered names, trademarks, etc. used in this book, even when not specifically marked as such, are not to be considered unprotected by law. Printed in the Federal Republic of Germany Printed on acid-free paper Typesetting Asco Typesetters, Hong Kong Printing betz-druck GmbH, Darmstadt Binding Litges & Dopf GmbH, Heppenheim Wiley Bicentennial Logo Richard J. Pacifico ISBN 978-3-527-31806-3

V

Contents Preface Foreword

XVII XIX

List of Contributors

XXI

Part I

Bone 1

1

Mineralization of Bone: An Active or Passive Process? Thorsten Schinke and Michael Amling

1.1 1.2 1.3 1.4 1.5 1.6

2

2.1 2.2 2.3 2.4 2.5 2.6

3

Abstract 3 Physiological and Pathological Mineralization 3 Inhibitors of Pathological Mineralization 5 Activators of Physiological Mineralization 6 The Key Role of Pyrophosphate 9 The Mysterious Role of the Endopeptidase Phex 11 Concluding Remarks 13 References 14 Bone Morphogenetic Proteins 19 Walter Sebald, Joachim Nickel, Axel Seher, and Thomas D. Mu¨ller

Abstract 19 Introduction 19 What is a Bone Morphogenetic Protein? 21 BMP Receptors are Composed of Diverse Type I and Type II Receptor Chains 23 The Basic Signaling Mechanism is the Same for BMPs and other TGF-b-like Proteins 24 Biochemistry and Cell Biology of Receptor Specificity 25 Structural Basis for Specificity and Affinity of BMP Receptor Interaction 27


VI

Contents

2.7

What We Can Do with BMPs: The Engineering of BMP-2 and GDF-5 Variants 29 References 32

3

Biomechanics of Bones: Modeling and Computation of Bone Remodeling 35 Udo Nackenhorst

3.1 3.2 3.3 3.3.1 3.3.2 3.3.3 3.3.4 3.4

4

4.1 4.2 4.2.1 4.2.2 4.3 4.3.1 4.3.2 4.3.3 4.4 4.5

5

5.1 5.1.1 5.1.1.1 5.1.1.2

Abstract 35 Introduction 35 The Biomechanical Equilibrium Approach 36 A Computational Multi-Scale Approach for Cortical Bone 42 Closed Nano-to-Meso Control Circuit Approach 43 Sub-Cellular Length-Scale 44 Micro-Scale Model (Single Osteon) 45 Meso-Scale Model of Cortical Bone 45 Conclusions 46 References 47 Direct X-Ray Scattering Measurement of Internal Stresses and Strains in Loaded Bones 49 Stuart R. Stock and Jonathan D. Almer

Abstract 49 Introduction 49 Background 50 X-Ray Scattering 50 Strains and Stresses 51 Methods 51 Specimens and Geometry 51 Analysis of Two-Dimensional (2-D) Scattering Patterns 53 X-Ray Elastic Constants and Strain–Stress Conversion 55 Examples of Data and Analysis 55 Discussion and Future Directions 56 References 57 Osteoporosis and Osteopetrosis Adele L. Boskey

59

Abstract 59 Introduction: Two Distinct Diseases with Common Features Comparisons of Clinical Features of Osteoporosis and Osteopetrosis 60 Histology 60 Radiography 61

59

Contents

5.1.2 5.2 5.2.1 5.2.1.1 5.2.1.2 5.2.2 5.2.2.1 5.2.2.2 5.3 5.3.1 5.3.2 5.4

6

6.1 6.2 6.3 6.4 6.4.1 6.4.2 6.5 6.6 6.7 6.8 6.9 6.10

7

7.1 7.2 7.3 7.4

Comparisons of Bone Mineral Properties in Osteoporosis and Osteopetrosis 62 Animal Models of Osteoporosis and Osteopetrosis 63 Osteoporosis 63 Rodent Models 63 Non-Rodent Models 66 Osteopetrosis 66 Rodent Models 67 Other Osteopetrotic Models 68 The Cellular and Molecular Bases of Osteopetrosis and Osteoporosis 70 Osteoporosis 70 Osteopetrosis 72 Biomineralization in Osteopetrosis and Osteoporosis 74 References 75 Biomimetic Bone Substitution Materials Matthias Epple

81

Abstract 81 The Clinical Need for Bone Substitution Materials 81 Synthetic Materials Used for Bone Substitution 82 Ceramics and Bone Cements 84 Polymers 86 PMMA-Based Materials 86 Polyester-Based Materials 86 Metals 87 Composites 87 Bone Substitutes of Biological Origin 87 Biological Functionalization of Synthetic Materials 90 An Example of a Synthetic Biomimetic Bone Substitution Material 90 Conclusions and Future Developments 91 References 92 Simulated Body Fluid (SBF) as a Standard Tool to Test the Bioactivity of Implants 97 Tadashi Kokubo and Hiroaki Takadama

Abstract 97 Introduction 97 Qualitative Correlation of Bone-Bonding Bioactivity of a Material with Apatite Formation on its Surface in SBF 98 Quantitative Correlation of Bone-Bonding Bioactivity and ApatiteForming Ability in SBF 100 Ion Concentrations of SBF 101

VII

VIII

Contents

7.5 7.6 7.7 7.8 7.9

Materials Able to Form Apatite 102 Composition and Structure of Apatite 103 Mechanism of Bonding of Bioactive Material to Bone 104 Mechanisms of Apatite Formation 105 Summary 106 References 106

8

Stimulation of Bone Growth on Implants by Integrin Ligands Mońica Lo´pez-Garcıá and Horst Kessler

8.1 8.1.1 8.1.2 8.1.3 8.1.4 8.2 8.2.1 8.2.2 8.2.3 8.2.4 8.2.4.1 8.2.4.2 8.2.4.3 8.2.4.4 8.3

9

9.1 9.2 9.2.1 9.2.2 9.2.3 9.2.4 9.2.5

109

Abstract 109 Introduction 109 Biomimetic Materials for Implant Technology 109 Integrins and RGD Sequence 110 Natural Proteins or Synthetic Peptides as Cell-Adhesive Molecules? 111 Integrin-Mediated Cell Adhesion 112 Improvements in Implant-Osseointegration by Surface Modification with Integrin Ligands 115 Mechanisms of Bone Grafting 115 Modifications on Implant Surfaces to Improve its Osseointegration 116 Structure of the Coating Molecules 117 Stimulation of Osteoblasts Adhesion and Proliferation on Implants Promoted by Integrin Ligands 118 Poly(methyl methacrylate) 118 Silks 120 Titanium 120 RGD Mimetics 121 Conclusions 123 References 123 Biochemical and Pathological Responses of Cells and Tissue to Micro- and Nanoparticles from Titanium and other Materials 127 Fumio Watari, Kazuchika Tamura, Atsruro Yokoyama, Kenichiro Shibata, Tsukasa Akasaka, Bunshi Fugetsu, Kiyotaka Asakura, Motohiro Uo, Yasunori Totsuka, Yoshinori Sato, and Kazuyuki Tohji

Abstract 127 Introduction 128 Materials and Methods 128 Specimens 128 Dissolution Testing of Ti Particles 129 Probe Cells 129 Biochemical Analyses of Cellular Reactions to Materials Animal Experiments 129

129

Contents

9.3 9.3.1 9.3.2 9.3.2.1 9.3.2.2 9.3.2.3 9.3.2.4 9.3.2.5 9.3.3 9.3.4 9.3.5 9.3.6 9.3.6.1 9.3.6.2 9.3.6.3 9.4 9.4.1 9.4.2 9.4.3 9.4.4 9.4.5

10

10.1 10.2 10.2.1 10.2.2 10.2.3 10.2.4 10.2.5 10.2.6 10.2.7 10.2.7.1 10.2.7.2 10.3 10.3.1 10.3.2

Results 130 Dependence of Tissue Reaction In Vivo on Material Macroscopic Size 130 Effect of Particle Size on Biocompatibility 130 Size Distribution of the Abraded Particles 130 Particle Size Dependence In Vitro 131 Particle Size Dependence In Vivo 133 Material Dependence of the Particle Size Effect In Vitro 135 Material Dependency of Tissue Reaction to Particles In Vivo 135 Shape Effect 136 The Origin of the Particle Size Effect 137 Toxicity Level of Particle Size Effect for Bioactive and Bioinert Materials 138 Nanotoxicology 139 Size-Dependent Stimulus Down to Nanometer Size 139 Internal Diffusion of Nanoparticles 139 Toxicity-Enhancing Effects of Biostimulatory Materials by Nanosizing 140 Discussion 140 Particle Size-Dependence of Cytotoxicity 140 Particle Size-Dependence in Soft Tissues 141 Comparison of Ti, Fe, and Ni Particles 141 The Effect of Micro-/Nanosizing on Biological Reactions 142 Terminology on ‘‘Nanotoxicology’’ 143 References 143 Tissue Engineering of Bone 145 Hans-Peter Wiesmann, Beate Lu¨ttenberg, and Ulrich Meyer

Abstract 145 Tissue Engineering: What Does it Mean? 145 Components of Bone Tissue Engineering 147 Osteoblasts 147 Bone Marrow Cells 148 Marrow-Derived Stem Cells 148 Vascular Cells 149 Scaffold Design and Cell Compatibility 149 Bioreactors 150 In-Vitro Cell Stimulation 150 Biophysical Stimulation 150 Biochemical Stimulation 151 Bone Biomineralization in Tissue Engineering Ex Vivo and In Vivo 151 Principles of ECM Biomineralization 151 Principles of Bone Formation 152

IX

X

Contents

10.3.3 10.4 10.5

Particular Features of Extracorporeal Biomineralization Clinical Demands 153 Future Aspects 154 References 155

Part II

Teeth

11

Formation of Teeth 159 Katharina Reichenmiller and Christian Klein

11.1 11.2 11.2.1 11.2.2 11.3 11.3.1 11.3.2 11.3.3 11.4 11.5 11.5.1 11.5.2 11.5.3 11.5.4 11.6

12

12.1 12.2 12.3 12.4 12.5

13

13.1 13.2

153

157

Abstract 159 Introduction 159 Odontogenesis 163 Genes Involved in Odontogenesis 165 Stem Cells 165 Dentinogenesis 165 Mantle and Circumpulpal Dentin 166 Intertubular Dentin 168 Peritubular Dentin 168 Amelogenesis 170 Cementogenesis 172 Acellular Extrinsic Fiber Cementum (AEFC) 172 Cellular Intrinsic Fiber Cementum (CIFC) 173 Cellular Mixed Stratified Cementum (CMSC) 174 Acellular Intrinsic Fiber Cementum (AIFC) 174 Acknowledgments 174 References 174 The Structure of Teeth: Human Enamel Crystal Structure Fre´de´ric Cuisinier and Colin Robinson

177

Abstract 177 Introduction 177 HRTEM Observations 178 AFM Observations 179 Discussion 181 Conclusions 182 References 182 Design Strategies of Human Teeth: Biomechanical Adaptations Paul Zaslansky and Steve Weiner

Abstract 183 Introduction 183 Deformation of Whole Teeth under Load

185

183

Contents

13.3 13.4 13.5 13.6

Mechanical Behavior of the Enamel Cap 191 The Role of Crown Dentin During Load Bearing 194 The Role of the Root and Supporting Structures 196 Broader Implications and Conclusions 198 References 200

14

Clinical Aspects of Tooth Diseases and their Treatment Peter Ga¨ngler and Wolfgang H. Arnold

14.1 14.2 14.2.1 14.2.2 14.3 14.4 14.5 14.5.1 14.5.2

15

15.1 15.2 15.3 15.4 15.4.1 15.4.2 15.5 15.5.1 15.5.2

16

16.1 16.2 16.2.1 16.2.2

203

Abstract 203 Introduction 203 Tooth Development and Developmental Anomalies 206 Developmental Features and Elemental Analysis of Early Mineralization 207 Developmental Anomalies 210 Dental Caries 212 Periodontal Diseases 216 Dental Trauma 220 Acute Dental Trauma 220 Chronic Dental Trauma 220 References 221 Dental Caries: Quantifying Mineral Changes Susan M. Higham and Philip W. Smith

223

Abstract 223 Introduction 223 Enamel Caries 224 Dentine Caries 224 Analyzing Mineral Changes in Dental Caries 225 Transverse Microradiography 226 TMR Studies 227 Quantitative Light-Induced Fluorescence 229 In Vitro QLF Studies 231 In Vivo QLF Studies 234 References 236 Periodontal Regeneration 239 Hom-Lay Wang and Lakshmi Boyapati

Abstract 239 Definitions 239 Periodontal Wound Healing 240 Wound-Healing Principles 240 Compartmentalization 241

XI

XII

Contents

16.2.3 16.3 16.3.1 16.3.1.1 16.3.2 16.3.2.1 16.3.2.2 16.3.2.3 16.3.2.4 16.3.3 16.3.3.1 16.3.3.2 16.3.4 16.3.4.1 16.3.4.2 16.3.4.3 16.3.4.4 16.4 16.4.1 16.4.2 16.4.3 16.4.4 16.5 16.5.1 16.5.2 16.5.3 16.5.4 16.6

Evaluating Regeneration 241 Techniques Used for Regeneration 241 Root Surface Biomodification 241 Root Surface Conditioning 241 Bone Replacement Grafts 242 Autografts 243 Allografts 243 Xenografts 245 Alloplasts 245 Guided Tissue Regeneration 246 Non-Absorbable Membranes 246 Absorbable Membranes 246 Biologic Modifiers 248 Growth Factors/Cytokines 248 Bone Morphogenetic Proteins (BMPs) 248 Pep-Gen p-15 249 Enamel Matrix Derivative (EMD) 249 Factors Influencing GTR Success 249 Patient Factors 251 Defect/Local Factors 251 Treatment Factors 252 Postoperative Care 252 Surgical Principles 253 Furcation Defects 253 Intrabony Defects 253 Root Coverage 253 Surgical Techniques 255 Conclusions 258 References 258

17

Tissue Engineering of Teeth Misako Nakashima

17.1 17.2 17.2.1 17.2.1.1 17.2.1.2 17.2.1.3 17.2.2 17.2.3 17.3 17.3.1

265

Abstract 265 Introduction 265 The Triad 266 Pulp Stem/Progenitor Cells 266 Isolation 266 Self-Renewal 268 Multipotential Differentiation 269 Morphogenetic Signals, BMPs 269 Scaffold 270 Dentin Regeneration 271 Protein Therapy 271

Contents

17.3.2 17.3.2.1 17.3.2.2 17.4 17.4.1 17.4.2 17.5 17.6

Gene Therapy 272 In-Vivo BMP Gene Therapy 272 Ex-Vivo BMP Cell Therapy and Gene Therapy 273 Pulp Regeneration 276 Vasculogenesis 276 Neurogenesis 276 Whole-Teeth Regeneration 277 Conclusions and Future Perspectives 278 References 278

Part III

Pathological Calcifications

18

Aspects of Pathological Calcifications Inge Schmitz

18.1 18.1.1 18.1.2 18.2 18.2.1 18.2.2 18.2.2.1 18.3 18.3.1 18.3.1.1 18.3.1.2 18.3.2 18.3.3 18.4 18.4.1 18.4.2 18.5

19

19.1 19.2 19.2.1 19.2.2 19.2.3

283 285

Abstract 285 Introduction 285 Examples of Pathological Calcification 286 Regulation of Calcifications 287 Heterotopic Ossification 288 Calcification in Ulcera of Patients with Paraplegia 288 Calcifications of the Lung 289 Metastatic Pulmonary Calcifications 291 Vascular Calcifications: Arteriosclerosis 291 Calcifications of Arteries 291 Calcification of the Tunica Media (Mo¨nckeberg’s Arteriosclerosis) 292 Calcification of the Tunica Intima (Arteriosclerosis) 293 Ossifications of Arteries 294 Characterization of Atherosclerotic Plaques of the Human Aorta 294 Calcification of Synthetic Vascular Grafts 296 Chronic Kidney Disease-Dialysis and Vascular Calcification of Arteries and Arteriovenous Shunts 296 Ossification of Synthetic Grafts 298 Conclusions 299 References 299 Atherosclerosis: Cellular Aspects 301 Diane Proudfoot and Catherine M. Shanahan

Abstract 301 Introduction 301 Role of VSMCs in Vascular Calcification 303 Release of Apoptotic Bodies and Vesicles 303 Phagocytosis 305 VSMC Osteo/Chondrocytic Conversion 306

XIII

XIV

Contents

19.2.4 19.3 19.3.1 19.3.2 19.4 19.5

20

20.1 20.2 20.3 20.4

21

Role of Calcifying Vascular Cells and Pericytes 309 Role of Inflammatory Cells 310 Macrophages 310 Dendritic Cells, Mast Cells and T Lymphocytes 312 The Role of Osteoclasts: Is there a Possibility for CalcificationRegression? 312 Conclusions 313 References 313 The Biological and Cellular Role of Fetuin Family Proteins in Biomineralization 317 Cora Scha¨fer and Willi Jahnen-Dechent

Abstract 317 Osteogenesis and Bone Mineralization versus Calcification 317 a2 -HS Glycoprotein/Fetuin-A is a Systemic Inhibitor of Ectopic Calcification 320 The Mechanism of Fetuin-A Inhibition of Calcification 322 The Fate of Calciprotein Particles 322 References 325 Stone Formation 329 Pierfrancesco Bassi

Abstract 329 Urinary Stones 329 Pathogenesis 330 Inhibitors of Stone Formation 331 Classification of Urinary Stones 333 Calcium Stones 333 Uric Acid Stones 336 Magnesium Ammonium Phosphate Stones, Struvite or Infection Stones 337 21.1.2.4 Cystine Stones 338 21.1.3 Risk Factors 338 21.1.3.1 Non-Genetic Factors 338 21.1.3.2 Genetic Factors 341 21.2 Other Urological Stones: Testicular Microlithiasis 343 21.3 Biliary and Gallbladder Stones 343 21.4 Miscellaneous Stones 344 21.4.1 Sialolithiasis 344 21.4.2 Dental Stones 344 21.4.3 Pancreatic Stones 345 21.4.4 Broncholithiasis and Pulmonary Alveolar Microlithiasis 345 References 346 21.1 21.1.1 21.1.1.1 21.1.2 21.1.2.1 21.1.2.2 21.1.2.3

Contents

22

Ectopic Mineralization: New Concepts in Etiology and Regulation Cecilia M. Giachelli

349

Abstract 349 Introduction 349 Regulators of Ectopic Mineralization 350 Circulating Factors that Regulate Ectopic Mineralization 350 Ion Transporters and Homeostatic Enzymes that Regulate Ectopic Mineralization 352 22.2.2.1 Role of Sodium-Dependent Phosphate Co-Transporters in Ectopic Mineralization 353 22.2.3 Extracellular Matrix Molecules that Regulate Ectopic Mineralization 354 22.2.3.1 Role of Osteopontin in Ectopic Mineralization 355 22.2.4 Cell Signaling Pathways that Regulate Ectopic Mineralization 355 22.2.5 Roles of Cell Death and Bone Remodeling in Ectopic Mineralization 357 22.3 Conclusions and Implications for Therapeutic Control of Ectopic Mineralization 358 References 358 22.1 22.2 22.2.1 22.2.2

23

23.1 23.2 23.3 23.4 23.5 23.6 23.7 23.8

24

24.1 24.2 24.3 24.4 24.4.1

Pathological Calcification of Heart Valve Bioprostheses Birgit Glasmacher and Martin Krings

361

Abstract 361 Introduction 361 In-Vitro Calcification Models 364 Heart Valve Bioprostheses 364 Calcification Hypotheses and Study Design 364 Calcification Imaging Methods 365 Calcification Patterns 367 Description of Findings 369 Conclusions and Future Research 370 References 371 The Biomaterials Network (Biomat.net) as a Major Internet Resource for Biomaterials, Tissue Engineering and Biomineralization 373 Pedro L. Granja, Jose´ Paulo Pereira, and Ma´rio A. Barbosa

Abstract 373 The Internet as a Major Healthcare Resource 373 Impact of Biomaterials Science in Modern Society 375 Biomat.net as a Biomineralization Resource 376 The Biomaterials Network (Biomat.net) 383 An Overview 383

XV

XVI

Contents

24.4.2 24.4.3 24.4.4 24.4.4.1 24.4.4.2 24.4.4.3 24.4.4.4 24.4.4.5 24.4.4.6 24.4.4.7

Objectives 384 Team 385 Functionalities 386 Site Map 387 Membership 387 Links Lists 387 Directory of Researchers 387 Jobs 388 Newsletter 388 Endorsement of Scientific Meetings References 389 Index 391

389

XVII

Preface When I began my academic career in Hamburg, I worked on the solid-state chemistry of organic compounds. By chance, our group identified a solid-state reaction which led to a polymer (polyglycolide), a class of biodegradable polyesters which a literature survey showed to have many applications in clinical medicine. This finding led in turn to my first contacts with physicians and their ideas about implants, biomaterials, and bone. As a member of a graduate school on bone in the medical faculty, I learned that biomaterials usually depend on biomineralization – that is, the cellular action and similarity to the corresponding hard tissue. I also became aware for the first time of the great importance of pathological calcifications in our society. The process of biomineralization is based on the formation of inorganic crystals, and is strongly controlled by organic molecules which are themselves controlled by cells. As a chemist, I was familiar with the first part of this story, but I had much to learn about biomolecules, the different cell types, and also the clinical treatments of diseases. Fortunately, in Hamburg, as well as in the subsequent institutions where I worked – at Bochum and Essen – I met people from the field of medicine who were interested in conducting joint studies. When all parties had learned to communicate – due to the different vocabulary of their disciplines – many fruitful collaborations developed, and in this respect I am deeply indebted to my colleagues (listed in alphabetical order), Prof. M. Amling, Prof. G. Delling, Prof. S. A. Esenwein, Prof. H. Eufinger, Rof. P. Ga¨ngler, Prof. W. JahnenDechent, Dr. A. Klocke, Prof. M. Ko¨ller, Dr. P. Lanzer, Dr. W. Linhart, Prof. G. Muhr, Prof. J. M. Rueger, Prof. W. Ruether, Dr. I. Schmitz, and Dr. S. Weihe, from whom I learned much about the clinical aspects of biomineralization. Therefore, I was very excited when Edmund Baüerlein asked me jointly to edit a volume of the Handbook of Biomineralization. We quickly selected the three main topics which are relevant in medicine, namely bone, teeth, and pathological calcifications, and also were fortunate to find many competent authors from all over the world who agreed to contribute. Many aspects of biomineralization in medicine are highlighted in this book, and I sincerely hope that it will contribute to our understanding of this field of research, as it is not only of academic inter-


XVIII

Preface

est but also covers many general aspects of biomineralization. Indeed, it can be said that good health is dependent on correct biomineralization! Essen, Germany February 2007

Matthias Epple

XIX

Foreword During the past decade, biomaterials research has undergone tremendous change. Ten years ago, the first dialogues were made between physicians searching for artificial materials that could be used for regenerative therapies (for example, in bone surgery), and materials scientists offering a variety of structural and/or functional materials which originally were developed for engineering applications. Today, we can take advantage of the impressive advances in modern biology and biochemistry, and as a consequence we have the chance to follow completely new pathways for solving such problems. As will be shown in this volume, our current understanding of biomineralization opens up new approaches not only for physicians but also for materials scientists in many medical applications. The key here is provided by the exploration of genetically controlled mechanisms of the biomineralization of hard tissues. The activation and inhibition of biomineralization are two complementary processes which lead to such wonderful structures as bone or teeth, and based on the phenotypic analyses of several mouse models and various diseases causing calcification of soft tissues, we have learned that there is a variety of noncollagenous proteins that control this interplay of activation and inhibition. The following examples will indicate how this knowledge can be used for biomimetic implant development. As the overall age of our modern-day society continues to rise, bone diseases will become increasingly important, and consequently the successful treatment of the pathological mineralization of bone – for example, in the case of osteoporosis – represents one of the major challenges of the next decade. This general problem in bone research is discussed in Chapter 5, wherein it is clear that, based on an understanding of biomineralization in such pathological situations, new strategies could be derived for biomedical treatments as well as for new materials that may be used in the regenerative therapy of disturbed tissue. Today, although a large variety of bone substitution materials is applied on a practical basis, an evolutionary process can be foreseen in which a group of artificial engineering materials will be completed by more biologically functionalized materials. Here, one favored strategy is the stimulation of bone growth on implants, and today the development of scaffolds suited to the immobilization of


XX

Foreword

living bone cells has been boosted by the recent progress made in stem cell research. Another exciting feature relating to the biomineralization of bone and teeth is the formation of hierarchical or graded structures. This leads to basic questions concerning structural evolution on the mesoscopic scale. The guiding principle for understanding such structures is related to the biomechanical adaptation of living tissue. In Chapter 13, the author explains how, in the case of teeth, evolutionary pressure has led to the creation of highly optimized structures with respect to the complex mechanical loading situations in a living organism. Based on this theory, many attempts have been made to model such structures, whereby the models describe the formation of the mineralized tissue by combining cellular activity with acting mechanical stresses. However, the predictive power of such numerical simulations based on Finite Element codes remains limited. In particular, the uncertainty of the constitutive laws for materials behavior sets such restrictions, and consequently new experimental approaches are required that will allow the appropriate measurement of the properties of these materials on the micro- and mesoscale. Today, there are growing numbers of promising methods available to perform just this task, and this is demonstrated throughout this volume. As mentioned above, the unwanted pathological calcification of soft tissue or of vascular systems is an issue which is closely connected with the biomineralization of hard tissues. Today, we know that there are no significant differences between both phenomena, and thus another broad field of research activity is opening up with relevance for biological tissue characterization as well as for the development of artificial materials, for example in vascular prostheses. In the final part of this volume, we show that this phenomenon is also dominated by active cell-mediated processes and not by the simple precipitation of minerals at a given substrate. In summarizing, it can be said that the exciting interdisciplinary cooperation of biologists, biochemists, materials scientists, and physicians has led us to a challenging new research field of biomineralization, with wonderful perspectives for a better understanding of the beauty of the evolution of living organisms, whilst at the same time making significant contributions to human healthcare. Prof. Dr. rer. nat Wolfgang Pompe Technische Universita¨t Dresden Institut fu¨r Werkstoffwissenschaft Professur fu¨r Materialwissenschaft und Nanotechnik Dresden Germany

XXI

List of Contributors Tsukasa Akasaka Graduate School of Dental Medicine Hokkaido University Kita 13, Nishi 7 Kita-Ku, Sapporo Japan

Kiyotaka Asakura Hokkaidou University Catalysis Research Center Kita 21, Nishi 10 Kita-ku, Sapporo Japan

Jonathan D. Almer Building 431 Advanced Photon Source Argonne National Laboratory 9700 South Cass Ave. Argonne, IL 60439 USA

Ma´rio A. Barbosa BIOMATERIALS NETWORK (Biomat.net) Biomaterials Laboratory INEB (Instituto de Engenharia Biomedica) University of Porto 823, Rua di Campo Alegre 4150-180 Porto Portugal

Michael Amling Clinics of Trauma-, Hand-, and Reconstruction Surgery University Medical Center Hamburg-Eppendorf Martinistraße 52 20246 Hamburg Germany Wolfgang H. Arnold Abteilung fu¨r Konservierende Zahnheilkunde Fakulta¨t fu¨r Zahn-, Mund- und Kieferheilkunde Universita¨t Witten/Herdecke Alfred-Herrhausen-Straße 50 58448 Witten Germany

Pierfrancesco Bassi Universita` Cattolica del Sacro Cuore Policlinico Universitario ‘A. Gemelli’ Largo A. Gemelli 00168 Rome Italy Adele L. Boskey Hospital for Special Surgery Weill Medical College of Cornell University 535 East 70th Street New York, NY 10021 USA


XXII


Lakshmi Boyapati School of Dentistry Department of Periodontics and Oral Medicine University of Michigan 1011 North University Avenue Ann Arbor MI 45109-1078 USA Fre´de´ric Cuisinier UFR Odontologie 545 Avenue du Professeur Jean-Louis Viala 34193 Montpellier Cedex 5 France Matthias Epple Institute of Inorganic Chemistry University of Duisburg-Essen Universita¨tsstraße 5–7 45117 Essen Germany Bunshi Fugetsu Graduate School of Environmental Science Hokkaido University Kita 10, Nishi 5 Kita-Ku, Sapporo Japan Peter Ga¨ngler Abteilung fu¨r Konservierende Zahnheilkunde Fakulta¨t fu¨r Zahn-, Mund- und Kieferheilkunde Universita¨t Witten/Herdecke Alfred-Herrhausen-Straße 50 58448 Witten Germany

Cecilia M. Giachelli Bioengineering Foege Hall Box 35506-1 University of Washington 1705 NE Pacific Street Seattle, WA 98195 USA Birgit Glasmacher Institute for Multiphase Processes and Center for Biomedical Engineering Gottfried Wilhelm Leibniz University Hannover Callinstraße 36 30167 Hannover Germany Pedro L. Granja BIOMATERIALS NETWORK (Biomat.net) Biomaterials Laboratory INEB (Instituto de Engenharia Biomedica) University of Porto 823, Rua di Campo Alegre 4150-180 Porto Portugal Susan M. Higham University of Liverpool Department of Clinical Dental Sciences Edwards Building Daulby Street Liverpool L69 3GN United Kingdom Willi Jahnen-Dechent Dept. of Biomedical Engineering Biointerface Laboratory RWTH Aachen University Hospital Pauwelsstraße 30 52074 Aachen Germany


Horst Kessler Department of Chemistry Institut fu¨r Organische Chemie und Biochemie Technische Universita¨t Mu¨nchen Lichtenbergstr. 4 85747 Garching Germany Christian Klein School of Dental Medicine Department of Operative Dentistry and Periodontology Osianderstraße 2–8 72076 Tu¨bingen Germany Tadashi Kokubo College of Life and Health Sciences Dept. of Biomedical Sciences Chubu University 1200 Matsumoto-cho Kasugai city, Aichi 487-8501 Japan Martin Krings Institute for Multiphase Processes and Center for Biomedical Engineering Gottfried Wilhelm Leibniz University Hannover Callinstraße 36 30167 Hannover Germany Mońica Lo´pez-Garcıá Department of Chemistry Institut fu¨r Organische Chemie und Biochemie Technische Universita¨t Mu¨nchen Lichtenbergstr. 4 85747 Garching Germany

¨ttenberg Beate Lu Department of Cranio-Maxillofacial Surgery University of Mu¨nster Waldeyerstr. 30 48149 Mu¨nster Germany Ulrich Meyer Clinic for Maxillofacial and Plastic Facial Surgery University of Du¨sseldorf Moorenstr. 5 40225 Du¨sseldorf Germany ¨ller Thomas Dieter Mu Universita¨t Wu¨rzburg, Biozentrum Physiologische Chemie II Am Hubland 97074 Wu¨rzburg Germany Udo Nackenhorst Institut fu¨r Baumechanik und Numerische Mechanik (IBNM) International Center for Computational Engineering Sciences (ICCES) Appelstraße 9A 30167 Hannover Germany Misako Nakashima Laboratory for Oral Disease Research National Institute for Longevity Sciences National Center of Geriatry and Gerontology 36-3 Gengo, Morioka, Obu Aichi 474-8522 Japan

XXIII

XXIV


Joachim Nickel Universita¨t Wu¨rzburg, Biozentrum Physiologische Chemie II Am Hubland 97074 Wu¨rzburg Germany Jose´ Paulo Pereira BIOMATERIALS NETWORK (Biomat.net) Biomaterials Laboratory INEB (Instituto de Engenharia Biomedica) University of Porto 823, Rua di Campo Alegre 4150-180 Porto Portugal Diane Proudfoot University of Cambridge Division of Cardiovascular Medicine ACCI Building Level 6, Box 110 Addenbrooke’s Hospital Hills Road Cambridge CB2 2QQ United Kingdom Katharina Reichenmiller School of Dental Medicine Department of Operative Dentistry and Periodontology Osianderstraße 2–8 72076 Tu¨bingen Germany Colin Robinson Leeds Dental Institute Clarendon Way Leeds LS29LU United Kingdom

Yoshinori Sato Graduate School of Environmental Studies Tohoku University 6-6-20, Aramaki Aza Aoba Aoba-ku, Sendai Miyagi Japan Cora Scha¨fer Dept. of Biomedical Engineering Biointerface Laboratory RWTH Aachen University Hospital Pauwelsstraße 30 52074 Aachen Germany Thorsten Schinke Clinics of Trauma-, Hand- and Reconstruction Surgery University Medical Center HamburgEppendorf Martinistraße 52 20246 Hamburg Germany Inge Schmitz Institute of Pathology and German Mesothelioma Register Ruhr University Bochum Bergmannsheil Clinic Buerkle-de-la-Camp-Platz 1 44789 Bochum Germany Walter Sebald Universita¨t Wu¨rzburg, Biozentrum Physiologische Chemie II Am Hubland 97074 Wu¨rzburg Germany


Axel Seher Universita¨t Wu¨rzburg, Biozentrum Physiologische Chemie II Am Hubland 97074 Wu¨rzburg Germany

Hiroaki Takadama College of Life and Health Sciences Deptartment of Biomedical Sciences Chubu University 1200 Matsumoto-cho Kasugaicity, Aichi 487-8501 Japan

Catherine M. Shanahan University of Cambridge Division of Cardiovascular Medicine ACCI Building Level 6, Box 110 Addenbrooke’s Hospital Hills Road Cambridge CB2 2QQ United Kingdom

Kazuchika Tamura Graduate School of Dental Medicine Hokkaido University Kita 13, Nishi 7 Kita-Ku, Sapporo Japan

Kenichiro Shibata Graduate School of Dental Medicine Hokkaido University Kita 13, Nishi 7 Kita-Ku, Sapporo Japan Philip W. Smith University of Liverpool Department of Dental Sciences Pembroke Place Liverpool L3 5PS United Kingdom Stuart R. Stock Department of Molecular Pharmacology and Biological Chemistry Mail code S-215 Northwestern University 303 East Chicago Avenue Chicago, IL 60611-3008 USA

Kazuyuki Tohji Graduate School of Environmental Studies Tohoku University 6-6-20, Aramaki Aza Aoba Aoba-ku, Sendai Miyagi Japan Yasunori Totsuka Graduate School of Dental Medicine Hokkaido University Kita 13, Nishi 7 Kita-Ku, Sapporo Japan Motohiro Uo Graduate School of Dental Medicine Hokkaido University Kita 13, Nishi 7 Kita-Ku, Sapporo Japan

XXV

XXVI


Hom-Lay Wang School of Dentistry Department of Periodontics and Oral Medicine University of Michigan 1011 North University Avenue Ann Arbor, MI 45109-1078 USA Fumio Watari Graduate School of Dental Medicine Hokkaido University Kita 13, Nishi 7 Kita-Ku, Sapporo Japan Steve Weiner Department of Structural Biology Faculty of Chemistry Weizmann Institute of Science Kimmelman Building (13) Rehovot 76100 Israel

Hans-Peter Wiesmann Department of Cranio-Maxillofacial Surgery University of Mu¨nster Waldeyerstr. 30 48149 Mu¨nster Germany Atsruro Yokoyama Graduate School of Dental Medicine Hokkaido University Kita 13, Nishi 7 Kita-Ku, Sapporo Japan Paul Zaslansky Dept. Biomaterials Max-Planck-Institute of Caloies and Interfaces Wissenschaftspark Golm Ann Mu¨hlenburg 1 14476 Golm Germany

Part I Bone


3

1 Mineralization of Bone: An Active or Passive Process? Thorsten Schinke and Michael Amling

Abstract

The most specific feature of bone tissue is the presence of a mineralized extracellular matrix (ECM) produced by osteoblasts. However, compared to our steadily increasing knowledge on the regulation of osteoblast proliferation and differentiation, our understanding of the underlying mechanisms which regulate bone ECM mineralization remain incomplete. Moreover, the phenotypic analysis of several mouse models and human patients with ectopic calcification of soft tissues has led to the hypothesis that bone mineralization might be rather a passive process, whereas ECM mineralization is actively inhibited in other tissues. Although this hypothesis is in line with the fact that extracellular concentrations of calcium and phosphate are far above the solubility product for spontaneous precipitation, there is accumulating evidence demonstrating that this view of ECM mineralization is rather simplified. This chapter provides a review of the most important findings from the genetic analysis of bone mineralization, using mouse models and human patients. Key words: Phex.

extracellular matrix, mineralization, osteoblast, Mgp, pyrophosphate,

1.1 Physiological and Pathological Mineralization

Physiological mineralization is restricted to bones, teeth and the hypertrophic zone of growth plate cartilage, whereas pathological mineralization – which more often is referred to as ‘‘ectopic calcification’’ – can be found in any tissue [1]. Whereas bone extracellular matrix (ECM) mineralization is the result of the cellular activity of osteoblasts, pathological ECM mineralization occurs mostly in the absence of osteoblasts, although there are some rare conditions where ectopic bone formation has been observed [2, 3]. With few exceptions (e.g., calcium oxaHandbook of Biomineralization. Edited by M. Epple and E. Baüerlein Copyright 8 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ISBN: 978-3-527-31806-3

4

1 Mineralization of Bone: An Active or Passive Process?

late crystals in nephrolithiasis), the mineral found not only in bone but also ectopically is composed mainly of calcium and phosphate, with a ratio indicative of crystalline hydroxyapatite [4, 5]. Thus, it is not surprising that low serum levels of calcium and/or phosphate can be one cause for defective skeletal mineralization [6, 7]. Likewise, elevated serum levels of calcium and/or phosphate are associated with a higher risk of ectopic calcification [8]. There is, however, increasing evidence that the so-called Ca P product is not the only factor influencing ECM mineralization, but that several gene products are involved in the local control of ECM mineralization in various tissues [9]. The importance of understanding the molecular mechanisms underlying ECM mineralization is underscored by the high prevalence of diseases associated with ectopic calcification or defects of skeletal mineralization. These include renal failure and atherosclerosis, where accompanying vascular calcifications are frequently observed, osteoarthritis, where mineral deposition occurs in the joints, as well as several inherited diseases, where the functional inactivation of specific genes leads to locally restricted pathological mineralization [10–14]. Likewise, defects of skeletal mineralization – such as rickets and osteomalacia – are not infrequent, and there is increasing evidence that not only the bone matrix volume but also the degree of bone matrix mineralization has a significant impact on the mechanical stability of the skeleton [15, 16]. In addition to the apparent clinical relevance of mineralization-related research, there is also an important biological question that needs to be addressed, as it is still virtually unknown why, under physiological conditions, ECM mineralization is restricted to the skeleton. In a simplified view, there are two possible explanations for this specificity: Only osteoblasts, odontoblasts and hypertrophic chondrocytes could produce factors inducing ECM mineralization – that is, skeletal mineralization is actively promoted. Only non-skeletal cell types could produce factors preventing ECM mineralization; that is, pathological mineralization is actively inhibited, whereas physiological mineralization is explained by the absence of inhibition. Although general opinion, which is reflected by several textbooks chapters on bone mineralization, was rather postulating an active mechanism, there is increasing evidence in favor of the second possibility. First, the extracellular concentrations of calcium and phosphate exceed their solubility product by several orders of magnitude, thus making all extracellular fluids metastable solutions in terms of calcium phosphate precipitation and mineral formation [17]. Second, several genes encoding inhibitors of pathological mineralization have been identified, based on the finding that their functional inactivation in mice and/or humans results in ectopic calcification [12–14, 18, 19]. Third, there is increasing evidence that tissue-non-specific alkaline phosphatase – one of the few gene products required for physiological mineralization of the bone ECM – acts by the removal of pyrophosphate, an inhibitor of mineral forma-

1.2 Inhibitors of Pathological Mineralization

tion [7, 20, 21]. However – and not surprisingly – biology is more complex, and there are still many observations that cannot be fitted into such a simplified concept of ECM mineralization. Therefore, at this point it is worth summarizing our current knowledge on ECM mineralization, and highlighting some of the questions that remain to be clarified by future experiments.

1.2 Inhibitors of Pathological Mineralization

Our current view of ECM mineralization was probably most influenced by the phenotypic analysis of two mouse models published ten years ago, where the genes encoding the two skeletal g-carboxyglutamate-containing proteins were specifically deleted [18, 22]. g-Carboxyglutamate (Gla) residues, resulting from a vitamin K-dependent post-translational modification step, are well known from factors of the blood coagulation system, where they increase binding to calciumloaded, negatively charged phospholipid surfaces [23]. Gla residues were also found in two proteins of the skeletal ECM, osteocalcin (also called bone Gla protein) and Mgp (matrix Gla protein). Whereas osteocalcin is specifically expressed by osteoblasts and represents a major constituent of the bone ECM, Mgp is expressed by growth plate chondrocytes, and also by vascular smooth muscle cells [18]. In both cases the negatively charged Gla residues are responsible for a highaffinity binding to hydroxyapatite, which made both osteocalcin and Mgp excellent candidates for a regulation of physiological ECM mineralization [24, 25]. Thus, it was very surprising that mice lacking osteocalcin did not display any defect of bone matrix mineralization, although their increased osteoblast activity resulted in a high bone mass phenotype [22]. In contrast, Mgp-deficient mice did have a mineralization defect of growth plate cartilage, though unexpectedly ECM mineralization was not diminished but extended into the prehypertrophic zone [18]. Moreover, the requirement of Mgp to inhibit ECM mineralization was much more obvious in the vasculature. In fact, all Mgp-deficient mice died around the age of 6 to 8 weeks due to rupture of the aorta, which became, like other major arteries, completely calcified over time [18]. Taken together, these results showed that the presence of Mgp in the ECM of the arterial wall, and also in the ECM of prehypertrophic cartilage, is required to prevent pathological mineralization. This is true not only for mice, but also for humans, where the functional inactivation of MGP in patients with the Keutel syndrome leads to similar phenotypic abnormalities [26]. In the meanwhile, the properties of the two skeletal Gla proteins were further compared using several transgenic mouse models [27]. While the presence of a transgene restoring Mgp expression in vascular smooth muscle cells completely rescued the vascular calcification of Mgp-deficient mice, the same approach with osteocalcin did not influence their phenotype at all. Moreover, while overexpression of Osteocalcin in osteoblasts had no influence on bone matrix mineralization, the same approach with Mgp led to severe defects, namely that the amount of

5

6


non-mineralized bone matrix (osteoid) was increased more than 10-fold [27]. These results demonstrate that only Mgp, and not Osteocalcin, has an inhibitory influence on ECM mineralization. Moreover, as Mgp is not a component of the bone ECM, and since its ectopic expression in osteoblasts interferes with bone mineralization, it must be concluded that the absence of mineralization inhibitors (e.g., Mgp) from the bone matrix is one prerequisite for physiological mineralization. The importance of Mgp in arteries and prehypertrophic cartilage is only one example of the need for specific gene products to prevent ectopic calcification. A second good example is the serum protein a2-HS-glycoprotein (Ahsg), also known as fetuin-A. As discussed in Chapter 22 of this volume, Ahsg is a systemically acting inhibitor of pathological mineralization, as the deficiency of Ahsg in calcification-prone mice leads to severe ectopic calcification of several organs, including kidney, lung, and skin [19]. Like Mgp, Ahsg has a high affinity for hydroxyapatite, and both proteins have been shown to be components of a high molecular-weight complex containing calcium and phosphate that was initially found in the serum of rats treated with etidronate [28]. In-vitro studies have further demonstrated that Ahsg acts by forming soluble colloidal spheres with calcium and phosphate in a manner comparable to the action of apolipoproteins that are required to solubilize lipids [29]. Besides these two examples, there is in fact further genetic evidence of the need for specific inhibitors of unwanted mineralization, although their mechanism of action is less well understood. Osteoprotegerin, for example, is a tumor necrosis factor (TNF)-receptor-like molecule which inhibits bone resorption, yet its deficiency in mice also results in arterial calcification [30]. In contrast, pseudoxanthoma elasticum – an inherited human disorder that is characterized by progressive calcification of the skin and blood vessels – is caused by a deficiency of the ATP-binding transmembrane protein ABCC6 through mechanisms that are still not fully understood [13]. Two other inhibitors of pathological mineralization, Ank and Enpp1, the deficiency of which causes ectopic calcification in mice and humans, act by raising the extracellular level of pyrophosphate, an inhibitor of ECM mineralization (this is discussed in detail below) [14, 31]. Regardless of the underlying mechanisms, however, there is indeed one obvious conclusion that must be drawn from this overwhelming genetic evidence, namely that pathological ECM mineralization needs to be prevented by active mechanisms – that is, by the expression of specific genes. Such a statement is completely in line with the astonishing fact that extracellular concentrations of calcium and phosphate are far above their solubility product; this means that, in theory, every organ should be calcified.

1.3 Activators of Physiological Mineralization

Compared to the number of mouse models and human diseases associated with pathological mineralization, there are fewer examples for genes that have been

1.3 Activators of Physiological Mineralization

Fig. 1.1 Three examples of skeletal mineralization defects. Undecalcified sections from the tibia of Vdr-deficient mice (left), Phex-deficient Hyp mice (middle), and from a bone biopsy of a TNAP-deficient hypophosphatasia patient (right) were stained using the von Kossa/van Gieson technique. Mineralized matrix is stained black; non-mineralized bone matrix (osteoid) is stained red. All three genetic defects cause a pathological enrichment of osteoid.

shown to be required for physiological mineralization of the skeleton. The most established ones encode the vitamin D receptor (Vdr), tissue-non-specific alkaline phosphatase (Tnap), and a bone-specific endopeptidase named Phex [32–34] (Fig. 1.1). Surprisingly, none of these proteins is a component of the bone ECM, and two of these genes (Vdr and Tnap) are not even expressed in a bone-specific fashion – which raises the question of how they can contribute to bone-specific ECM mineralization. In the case of Tnap and Phex, the molecular mechanisms underlying their functions are still being investigated, and are discussed below. In contrast, the requirement for Vdr in bone mineralization can already be explained, based mainly on the analysis of a mouse model with a targeted deletion of the Vdr gene. Vdr-deficient mice display all the characteristics of vitamin D-dependent rickets, including alopecia, as known from human patients with inactivating VDR mutations [35]. In addition to the striking defect of growth plate calcification in Vdrdeficient mice, their bone-specific histomorphometric analysis further revealed an osteoidosis; that is, an enrichment of non-mineralized bone matrix [36]. In order to provide an explanation for this phenotype, we utilized this mouse model and fed them a high-calcium diet, thereby fully normalizing their hypocalcemia, which is another phenotypic aspect of Vdr-deficiency, in mice and humans. Following histologic and histomorphometric analysis of these non-hypocalcemic Vdr-deficient mice, no defects of cartilage and bone ECM mineralization were observed; equally important, no changes in any of the parameters of bone formation and bone resorption were observed [36]. Thus, unlike the situation in alopecia, the severe mineralization defect associated with Vdr-deficiency is fully rescued by normalizing the serum calcium levels. This in turn indicates that the major physiologically relevant action of Vdr lies in the intestinal uptake of calcium, and not in any direct stimulation of bone ECM mineralization.

7

8


These findings were not especially unexpected, and did not challenge the classic concept that bone ECM mineralization is actively promoted. In fact, it was still possible that the bone ECM contains proteins that are specifically expressed by osteoblasts, and are needed to induce the formation of hydroxyapatite crystals. The major protein component of the bone matrix is type I collagen. As it has been shown that mutations in one of the two genes encoding this heterotrimeric protein are the cause of osteogenesis imperfecta, there is no doubt that type I collagen provides a molecular scaffold that is important for the ordered deposition of hydroxyapatite [37]. However, the presence of type I collagen in the bone ECM cannot by itself explain the specificity of the mineralization process, especially as it is also expressed by fibroblasts. Moreover, two recently described transgenic mouse models with ectopic Tnap expression provided evidence that ECM mineralization can be induced in the skin, when Tnap is expressed in fibroblasts, but not when it is expressed in keratinocytes [7]. Taken together, these findings demonstrate that type I collagen is required, but is not sufficient for ECM mineralization. Other major protein components of the bone ECM include the above-described Ahsg, which is enriched from the serum, but not expressed by osteoblasts [38], osteocalcin, osteopontin and bone sialoprotein (Bsp); the latter three components are strongly, if not specifically, expressed by osteoblasts [39]. As discussed above, neither the deficiency nor the overproduction of osteocalcin in transgenic mice has any influence on mineralization of the bone ECM. In fact, although osteocalcin-deficient mice display a high bone mass phenotype, their increased amount of bone matrix is normally mineralized [22]. Neither do osteopontindeficient mice display any defects of bone ECM mineralization [40]. Moreover, a combination of osteopontin- and Mgp-deficiency in mice leads to a further enhancement of the vascular calcification observed in the Mgp-deficient mice alone, demonstrating that osteopontin rather acts as an inhibitor of ECM mineralization, and not as an activator [41]. Among the known protein components of the mineralized bone ECM, Bsp was always considered to be the best candidate for an activator of ECM mineralization, as it is able to promote hydroxyapatite formation in vitro [42]. Thus, it was especially surprising that the targeted deletion of the Bsp gene in mice also does not result in obvious defects of skeletal mineralization [43]. Therefore, until recently, the only ECM protein that has been shown to influence mineralization in vivo was Mgp. Moreover, given the striking ectopic mineralization of the Mgpdeficient mice and the ability of Mgp to interfere with bone ECM mineralization when ectopically expressed in osteoblasts, it had to be concluded that there is no need for specific activators of physiological mineralization residing in the bone ECM [27]. However, as the functions of an increasing number of genes have been studied in vivo – mostly by their inactivation in mice – it was not too surprising that one bone ECM protein which is required for proper skeletal mineralization has finally been identified, namely dentin matrix protein 1 (Dmp1) [44]. As the name implies, Dmp1 was originally identified in teeth, but subsequently was also found to be expressed by differentiated osteoblasts [45, 46]. Dmp1, like Bsp and osteo-

1.4 The Key Role of Pyrophosphate

pontin, belongs to a family of integrin-binding acidic glycoproteins, and has been shown to induce hydroxyapatite formation in vitro [47]. Consistent with this activity, Dmp1-deficient mice display not only a severe hypomineralization of dentin but also a severe osteoidosis; that is, a pathological enrichment of non-mineralized bone matrix [44, 48]. However, even in the case of Dmp1-deficient mice, the explanation for the skeletal mineralization defect is not as simple as initially anticipated, as these mice have marked reductions in serum calcium and phosphate levels, for reasons that are still unknown [44]. It is not yet clear, therefore, whether it is the direct action of Dmp1 in the bone ECM that is required for its proper mineralization.

1.4 The Key Role of Pyrophosphate

As ECM proteins – with the exception of Mgp, and possibly Dmp1 – do not appear to play the most dominant roles in ECM mineralization, it is not surprising that other cellular activities participate in regulating physiological and pathological mineralization. One mechanism by which cells control ECM mineralization involves the regulation of extracellular pyrophosphate levels. The importance of such regulation has only recently been demonstrated by the functional analysis of three gene products, the deficiency of which in mice and humans leads to defects in physiological or pathological mineralization, namely Tnap, Enpp1, and Ank. There is hallmark evidence demonstrating that the activity of Tnap is required for skeletal mineralization, as several inactivating mutations within the human TNAP gene have been identified as the cause of hypophosphatasia, an inherited disorder characterized by defective mineralization of the bone ECM [33]. Likewise, Tnap-deficient mice recapitulate the phenotype of the human patients and develop severe osteoidosis [49]. Both, the patients with hypophosphatasia and the Tnap-deficient mice, have elevated serum levels of three phosphocompounds – phosphoethanolamine, pyridoxal phosphate and inorganic pyrophosphate – that therefore appear to be natural substrates for Tnap [49, 50]. In particular, the increased pyrophosphate levels (not only based on the genetic evidence discussed below) provided important information with regards to a possible mechanism for the physiologic action of Tnap. Inorganic pyrophosphate (PPi), which structurally is composed of two phosphate ions linked by an ester bond, has the ability to bind to nascent hydroxyapatite crystals, thereby preventing their further growth [51, 52]. Tnap, an ectoenzyme linked to the cell membrane via a GPI-anchor [53], has been shown to hydrolyze PPi, thereby lowering its extracellular concentration and producing phosphate ions (Pi) in turn [54, 55]. Modulation of the PPi/Pi ratio by Tnap appears to be the most important function of this enzyme, as the cell-autonomous mineralization defect of Tnap-deficient osteoblasts can be counteracted by the exogenous addition of Pi [7]. Likewise, lowering the PPi levels in Tnap-deficient osteoblasts by combining Tnap-deficiency with a deficiency of Enpp1 or Ank also rescues their mineralization defect, thus demonstrating that the PPi-hydrolyzing

9

10


activity of Tnap is indeed the predominant action, which explains its requirement for bone ECM mineralization [21, 56]. In contrast to Tnap, Enpp1 and Ank act by increasing the extracellular PPi concentration through different mechanisms. Enpp1 (also known as PC-1; plasma cell membrane glycoprotein 1) is a PPi-generating enzyme which resides in the cell membrane not only of osteoblasts but also of other cell types [57, 58]. The importance of Enpp1 in the regulation of physiological and pathological mineralization was first demonstrated by the genetic analysis of human patients suffering from OPLL (ossification of the posterior longitudinal ligament of spine), and a corresponding mouse model termed Ttw (tiptoe walking). In both cases, inactivating mutations within the Enpp1 gene cause ectopic calcification of the spinal ligaments, thereby providing another example of an inhibitor of pathological mineralization [31, 59]. Moreover, certain mutations of the human ENPP1 gene have recently been shown to cause IIAC (idiopathic infantile arterial calcification), which is another inherited disease with severe ectopic calcification of arteries [60]. As Enpp1-deficient mouse models have been shown to display increased bone formation and hypermineralization, it appears that the activity of Enpp1 is also required for the regulation of physiological mineralization [21]. Likewise, the importance of the Ank gene was first identified by the demonstration that its mutational inactivation in mice causes progressive ankylosis, characterized by ectopic calcification within the synovial fluid [61]. The Ank gene encodes a transmembrane protein which shuffles intracellular PPi to the extracellular space, thereby also leading to an increased PPi/Pi ratio in the ECM. ANK mutations in humans have been shown to cause either chondrocalcinosis, or craniometaphyseal dysplasia, the latter condition being characterized by hyperostosis of the calvarial and facial bones [14, 62]. Thus, like Enpp1, Ank is required for the regulation of pathological and physiological mineralization, where both proteins have an inhibitory action mediated through the elevation of extracellular PPi concentrations. In conclusion, hallmark genetic evidence is available showing that the extracellular PPi/Pi ratio plays a key role in ECM mineralization, and that at least three enzymes are physiologically involved in this regulation. As mentioned above, the best demonstration of the interactions of Tnap, Enpp1 and Ank came from the combination of the corresponding mouse deficiency models. This analysis demonstrated that the mineralization defect of the Tnap-deficient mice is rescued by an absence of the antagonistic actions of either Enpp1 or Ank [21, 56]. It is clear, therefore, that the requirement of Tnap for physiological mineralization can be explained by the removal of PPi, an inhibitor of mineral formation. Moreover, these findings also support the concept that ECM mineralization is mainly controlled by inhibitory mechanisms that must be released in the skeletal microenvironment. The question remains, however, of how these enzymes contribute to the tissuespecificity of the mineralization process, as they are all produced by osteoblasts, but not only. While Enpp1 and Ank are virtually expressed in ubiquitous fashion, there is at least some restriction in the case of Tnap, which is expressed in bone, kidney, liver, and testes [7]. Interestingly, the co-expression of Tnap with genes

1.5 The Mysterious Role of the Endopeptidase Phex

encoding type I collagen can only be found in bone – and this may even be the solution to the problem. As mentioned above, ectopic expression of Tnap in the skin of transgenic mice leads to ectopic mineralization only when the gene is coexpressed with type I collagen [7]. Thus, it is possible that physiological mineralization of the skeleton is due simply to the co-existence of a collagenous molecular scaffold and the activity of an enzyme that reduces the local levels of PPi. There is, however, at least one further mechanism clearly involved in regulating physiological mineralization, and this includes the activity of Phex.

1.5 The Mysterious Role of the Endopeptidase Phex

X-linked hypophosphatemic rickets (XLH) is the most common inherited disease in humans that is associated with defects of skeletal mineralization [63]. As the name implies, XLH patients also display low circulating phosphate levels that result from increased urinary phosphate excretion. The genetic defect underlying this phenotype was identified in 1995, and the affected gene subsequently termed Phex (phosphate-regulating gene with homologies to endopeptidases on the Xchromosome) [34]. Further analysis revealed that this gene encodes a transmembrane protein specifically expressed by osteoblasts (and not by kidney cells) with an extracellular zinc-binding domain sharing homologies with a family of endopeptidases. This family includes neutral endopeptidase (NEP), the endothelinconverting enzymes (ECE-1 and ECE-2) and the KELL antigen, which act by cleaving specific substrates such as substance P and enkephalin in the case of NEP, or the endothelin precursor protein in the case of ECE-1 and ECE-2 [64]. Thus, it appeared that the activity of Phex could lie in the activation or inactivation of osteoblast-produced substrate(s) regulating bone ECM mineralization and/or phosphate homeostasis. Until now, the major problem is that no physiological Phex substrate has yet been identified, although data are accumulating with regards to the pathophysiology of XLH. Most of this information has been derived from analyses of a Phex-deficient XLH mouse model (the Hyp mouse), but also from the genetic analysis of human patients with autosomal dominant hypophosphatemic rickets (ADHR) or tumor-induced osteomalacia (TIO) (see below). The Hyp mouse was established and already characterized as a model of XLH long before the demonstration that their genetic defect lies in an inactivating deletion of the Phex gene [65–67]. Several experiments using these mice led to the hypothesis that two substrates of Phex might exist – one regulating bone ECM mineralization in an autocrine/paracrine manner, and the other one regulating renal phosphate handling [68, 69]. As Phex is not expressed in the kidney, the latter substrate was postulated to be a circulating factor and hypothetically termed ‘‘phosphatonin’’. The existence of such a factor was confirmed by parabiosis experiments, where the surgical connection of the vasculature from wild-type and Hyp mice led to an increased urinary phosphate excretion of the parabiosed wild-type mice [70]. Moreover, cross-transplantation experiments showed that

11

12


transplanting the kidney from Hyp mice into wild-type mice did not affect phosphate homeostasis, whereas the opposite experiment led to hypophosphatemia [71]. Several biochemical experiments further demonstrated that the increased urinary phosphate excretion is caused by down-regulation of the expression of Npt2, a sodium-dependent phosphate transporter localized in the proximal tubule [72, 73]. In contrast to the pathogenesis of the hypophosphatemia, controversy persists regarding the question of whether the defects of skeletal mineralization in the absence of Phex are the result of decreased serum phosphate levels, or whether they are caused by cell-autonomous defects. In favor of the first possibility, it was recently reported that the severe osteoidosis observed in Hyp mice can almost be normalized by feeding them a high-phosphate diet [7]. There is, however, also convincing evidence in support of an intrinsic defect of bone ECM mineralization associated with the Phex-deficiency that is independent of the hypophosphatemia. First, transplanting bone cells from Hyp mice into the gluteal muscle of wild-type mice showed that their ability to form mineralized bone is reduced compared to transplanted bone cells from wild-type mice [74]. Second, at least some groups have found that primary osteoblast cultures derived from Hyp mice have a reduced ability to form mineralized bone nodules ex vivo [75]. Third – and probably most convincing – mice lacking Npt2, the renal phosphate transporter downregulated by phosphatonin, display hypophosphatemia, but not the characteristic defects of skeletal mineralization associated with deletion of the Phex gene [76]. Based on these findings, a second Phex substrate was postulated and termed ‘‘minhibin’’, as it is hypothetically involved in the local control of bone ECM mineralization. Whereas the identity of minhibin is still not clear, it appears that phosphatonin – the phosphaturic factor elevated in XLH – has already been identified as Fgf23 [69]. The importance of Fgf23 in the pathogenesis of XLH was first discovered by the parallel analysis of two human diseases sharing striking similarities to XLH, namely autosomal dominant hypophosphatemic rickets (ADHR) or tumorinduced osteomalacia (TIO). ADHR is an inherited disease, and its genetic analysis revealed that it is caused by mutations of the human FGF23 gene that lead to increased stability of the circulating protein [77, 78]. TIO is an acquired disorder, where certain tumors of mesenchymal origin lead to hypophosphatemia and defects of skeletal mineralization [79]. In a screen for genes that are differentially expressed in a bony tumor inducing osteomalacia, Fgf23 was found to be strongly up-regulated compared to the adjacent bone tissue [80]. Based on these findings, subsequent experiments were performed which showed that the injection of recombinant Fgf23 into wild-type mice leads to decreased renal phosphate reabsorption, and that the implantation of Fgf23-producing Chinese hamster ovary (CHO) cells into nude mice induces a phenotype reminiscent of XLH. Further experiments by several other groups showed that Fgf23 has the ability to down-regulate the expression of Npt2 in kidney cells, and that the elevation of Fgf23 serum levels correlates with the degree of hypophosphatemia in XLH patients [81–84]. Based on all this genetic and experimental evidence, it was specu-

1.6 Concluding Remarks

lated that Fgf23 is indeed the long-sought phosphatonin, which is physiologically inactivated by Phex-mediated cleavage and therefore enriched in the absence of functional Phex. Thus, several groups attempted to show that Fgf23 is indeed a Phex-substrate, but the final conclusion from these experiments was that Fgf23 is not cleaved by the endopeptidase activity of Phex [85, 86]. In contrast, the comparison of osteoblast cultures derived from wild-type and Hyp mice revealed an increased expression of Fgf23, suggesting that its up-regulation is caused by transcriptional mechanisms, and not by a difference in protein degradation. Thus, if Phex really acts as an endopeptidase, a physiological substrate still remains to be identified that can fill the ‘‘black box’’ between Phex-deficiency and an increased expression of the phosphaturic factor Fgf23. One good candidate for such a function was an ECM protein termed Mepe (matrix extracellular phosphoglycoprotein), which is specifically expressed in terminally differentiated osteoblasts and strongly upregulated in TIO tumors [87, 88]. However, like Fgf23, Mepe is not cleaved by the endopeptidase activity of Phex, and the combination of both deficiencies achieved by crossing Mepe-deficient mice with Hyp mice, did not alter the hypophosphatemia or the defects of skeletal mineralization observed in the latter animals [89, 90]. Thus, despite the efforts of many different laboratories, Phex is still an endopeptidase in search of a physiological substrate. However, there is no doubt that Fgf23 plays a major role in phosphate homeostasis, which was finally confirmed by the generation of a Fgf23-deficient mouse model. As expected, Fgf23-deficient mice display a hyperphosphatemia caused by increased renal phosphate reabsorption, thus demonstrating that the phosphaturic activity of Fgf23 is indeed of physiological importance [91]. In contrast, it was completely unexpected that these mice were severely growth-retarded and displayed a striking increase in the amount of osteoid, which was almost comparable to the skeletal mineralization defect observed in Hyp mice. The fact that both, the Fgf23-deficient mice and the Hyp mice display an osteoidosis – despite having oppositely regulated serum phosphate levels – underscores that the defects of skeletal mineralization cannot be simply explained by pathological alterations of phosphate homeostasis. Moreover, given the fact that both mouse models display a similar bone phenotype, which is still evident when both deficiencies are genetically combined [92], it must be concluded that Fgf23 can not be the only factor involved in the pathophysiology associated with the deficiency of Phex in mice and humans. Thus, it is very clear that there is at least one factor missing before the important role of this axis in bone ECM mineralization can be fully understood.

1.6 Concluding Remarks

Why does bone mineralize, and not every other tissue? This simple question can now be partially answered, although some open questions remain. It should be

13

14


pointed out that the ability to create genetically modified mouse models, and to pinpoint the genetic defects of inherited human disorders, were absolute requirements when answering the above question. Thus, our knowledge on physiological and pathological mineralization has begun to explode only during the past decade – which means that the final issues will most likely be solved very soon. These issues include the identification of those molecular mechanisms explaining the phenotypic abnormalities associated with the Phex-deficiency, as well as the interaction between the Phex/Fgf23 system and the enzymes regulating local concentrations of pyrophosphate. Concerning the role of the ECM proteins it remains to be clarified, whether Dmp1 is a direct activator of physiological mineralization, or whether the mineralization defects associated with its deficiency are caused by hypocalcemia. It is surprising that, to date, Mgp has been the only ECM protein shown to be required to inhibit ectopic calcification – which raises the question of whether the arterial wall is especially protected, or whether other tissues have their own ECM inhibitors of mineralization. Regardless of these remaining questions, it is already possible to draw a conclusion concerning the philosophical question, whether bone mineralization is an active or passive process. As the ectopic expression of Mgp in osteoblasts leads to a severe reduction in bone ECM mineralization, it is quite clear that an absence of potent mineralization inhibitors from the bone ECM contributes to its ability to mineralize. Moreover, as Tnap clearly acts by removing the mineralization inhibitor pyrophosphate, it appears that ECM mineralization is generally regulated by inhibitory mechanisms. Thus, it would not be too surprising, if the results of future experiments were to indicate that even the endopeptidase Phex acts by inactivating an inhibitor of mineralization. Although this latter issue is at present speculative, a current concept of ECM mineralization can already be established: it is actively inhibited in extraskeletal tissues, but this is counter-acted by specific gene products in the skeleton.

References 1 T. Schinke, M.D. McKee, G. Karsenty, 2

3

4 5

Nat. Genet. 1999, 21, 150–151. K. Bostrom, K.E. Watson, W.P. Stanford, L.L. Demer, Am. J. Cardiol. 1995, 75, 88B–91B. K.M. Galvin, M.J. Donovan, C.A. Lynch, R.I. Meyer, R.J. Paul, J.N. Lorenz, V. Fairchild-Huntress, K.L. Dixon, J.H. Dunmore, M.A. Gimbrone, Jr., D. Falb, D. Huszar, Nat. Genet. 2000, 24, 171–174. F.L. Coe, A. Evan, E. Worcester, J. Clin. Invest. 2005, 115, 2598–2608. A.S. Posner, Clin. Orthop. Relat. Res. 1985, 200, 87–99

6 Y.C. Li, M. Amling, A.E. Pirro,

M. Priemel, J. Meuse, R. Baron, G. Delling, M.B. Demay, Endocrinology 1998, 139, 4391– 4396. 7 M. Murshed, D. Harmey, J.L. Millan, M.D. McKee, G. Karsenty, Genes Dev. 2005, 19, 1093–1104. 8 G.A. Block, T.E. Hulbert-Shearon, N.W. Levin, F.K. Port, Am. J. Kidney Dis. 1998, 31, 607–617. 9 M. Ketteler, V. Brandenburg, W. Jahnen-Dechent, R. Westenfeld, J. Floege, Nephrol. Dial. Transplant. 2005, 20, 673–677.

References 10 W.G. Goodman, J. Goldin, B.D.

11 12

13

14

15 16 17

18

19

20

21

Kuizon, C. Yoon, B. Gales, D. Sider, Y. Wang, J. Chung, A. Emerick, L. Greaser, R.M. Elashoff, I.B. Salusky, N. Engl. J. Med. 2000, 342, 1478–1483. D. Hamerman, N. Engl. J. Med. 1989, 320, 1322–1330. F. Rutsch, N. Ruf, S. Vaingankar, M.R. Toliat, A. Suk, W. Hohne, G. Schauer, M. Lehmann, T. Roscioli, D. Schnabel, J.T. Epplen, A. Knisely, A. Superti-Furga, J. McGill, M. Filippone, A.R. Sinaiko, H. Vallance, B. Hinrichs, W. Smith, M. Ferre, R. Terkeltaub, P. Nurnberg, Nat. Genet. 2003, 34, 379–381. B. Struk, L. Cai, S. Zach, W. Ji, J. Chung, A. Lumsden, M. Stumm, M. Huber, L. Schaen, C.A. Kim, L.A. Goldsmith, D. Viljoen, L.E. Figuera, W. Fuchs, F. Munier, R. Ramesar, D. Hohl, R. Richards, K.H. Neldner, K. Lindpaintner, J. Mol. Med. 2000, 78, 282–286. A. Pendleton, M.D. Johnson, A. Hughes, K.A. Gurley, A.M. Ho, M. Doherty, J. Dixey, P. Gillet, D. Loeuille, R. McGrath, A. Reginato, R. Shiang, G. Wright, P. Netter, C. Williams, D.M. Kingsley, Am. J. Hum. Genet. 2002, 71, 933–940. B. Wharton, N. Bishop, Lancet 2003, 362, 1389–1400. G. Boivin, P.J. Meunier, Osteoporosis Int. 2003, 14, S19–S24. W.F. Neuman, in: M.R. Urist (Ed.), Fundamental and Clinical Bone Physiology. Lippincott Co., Philadelphia, 1980, pp. 83–107. G. Luo, P. Ducy, M.D. McKee, G.J. Pinero, E. Loyer, R.R. Behringer, G. Karsenty, Nature 1997, 386, 78–81. C. Schafer, A. Heiss, A. Schwarz, R. Westenfeld, M. Ketteler, J. Floege, W. Muller-Esterl, T. Schinke, W. JahnenDechent, J. Clin. Invest. 2003, 112, 357–366. R. Gijsbers, H. Ceulemans, W. Stalmans, M. Bollen, J. Biol. Chem. 2001, 276, 1361–1368. D. Harmey, L. Hessle, S. Narisawa, K.A. Johnson, R. Terkeltaub, J.L. Millan, Am. J. Pathol. 2004, 164, 1199–1209.

22 P. Ducy, C. Desbois, B. Boyce, G.

23 24 25 26

27

28

29

30

31

32

33 34 35

36

Pinero, B. Story, C. Dunstan, E. Smith, J. Bonadio, S. Goldstein, C. Gundberg, A. Bradley, G. Karsenty, Nature 1996, 382, 448–452. B. Furie, B.C ¸ . Furie, Ann. N. Y. Acad. Sci. 1991, 614, 1–10. Q.Q. Hoang, F. Sicheri, A.J. Howard, D.S. Yang, Nature 2003, 425, 977–980. M.E. Roy, S.K. Nishimoto, Bone 2002, 31, 296–302. P.B. Munroe, R.O. Olgunturk, J.P. Fryns, L. Van Maldergem, F. Ziereisen, B. Yuksel, R.M. Gardiner, E. Chung, Nat. Genet. 1999, 21, 142– 144. M. Murshed, T. Schinke, M.D. McKee, G. Karsenty, J. Cell. Biol. 2004, 165, 625–630. P.A. Price, G.R. Thomas, A.W. Pardini, W.F. Figueira, J.M. Caputo, M.K. Williamson, J. Biol. Chem. 2002, 277, 3926–3934. A. Heiss, A. DuChesne, B. Denecke, J. Grotzinger, K. Yamamoto, T. Renne, W. Jahnen-Dechent, J. Biol. Chem. 2003, 278, 13333–13341. N. Bucay, I. Sarosi, C.R. Dunstan, S. Morony, J. Tarpley, C. Capparelli, S. Scully, H.L. Tan, W. Xu, D.L. Lacey, W.J. Boyle, W.S. Simonet, Genes Dev. 1998, 12, 1260–1268. I. Nakamura, S. Ikegawa, A. Okawa, S. Okuda, Y. Koshizuka, H. Kawaguchi, K. Nakamura, T. Koyama, S. Goto, J. Toguchida, M. Matsushita, T. Ochi, K. Takaoka, Y. Nakamura, Hum. Genet. 1999, 104, 492–497. M.R. Hughes, P.J. Malloy, D.G. Kieback, R.A. Kesterson, J.W. Pike, D. Feldman, B.W. O’Malley, Science 1988, 242, 1702–1705. E. Mornet, Hum. Mutat. 2000, 15, 309–315. The HYP Consortium, Nat. Genet. 1995, 11, 130–136. Y.C. Li, A.E. Pirro, M. Amling, G. Delling, R. Baron, R. Bronson, M.B. Demay, Proc. Natl. Acad. Sci. USA 1997, 94, 9831–9835. M. Amling, M. Priemel, T. Holzmann, K. Chapin, J.M. Rueger, R. Baron, M.B. Demay, Endocrinology 1999, 140, 4982–4987.

15

16

1 Mineralization of Bone: An Active or Passive Process? 37 B. Sykes, D. Ogilvie, P. Wordsworth, 38

39 40

41

42

43

44

45

46

47 48

49

50 51 52 53

J. Anderson, N, Lancet 1986, 2, 69–72. T. Schinke, C. Amendt, A. Trindl, O. Poschke, W. Muller-Esterl, W. JahnenDechent, J. Biol. Chem. 1996, 271, 20789–10796. J.E. Aubin, Biochem. Cell. Biol. 1998, 76, 899–910. S.R. Rittling, H.N. Matsumoto, M.D. McKee, A. Nanci, X.R. An, K.E. Novick, A.J. Kowalski, M. Noda, D.T. Denhardt, J. Bone Miner. Res. 1998, 13, 1101–1111. M.Y. Speer, M.D. McKee, R.E. Guldberg, L. Liaw, H.Y. Yang, E. Tung, G. Karsenty, C.M. Giachelli, J. Exp. Med. 2002, 196, 1047–1055. G.K. Hunter, H.A. Goldberg, Proc. Natl. Acad. Sci. USA 1993, 90, 8562– 8565. J. Aubin, A. Gupta, R. Zinrgbl, J. Rossant, J. Bone Miner. Res. 1996, 11, S102. Y. Ling, H.F. Rios, E.R. Myers, Y. Lu, J.Q. Feng, A.L. Boskey, J. Bone Miner. Res. 2005, 20, 2169–2177. A. George, B. Sabsay, P.A. Simonian, A. Veis, J. Biol. Chem. 1993, 268, 12624–12630. R.N. D’Souza, A. Cavender, G. Sunavala, J. Alvarez, T. Ohshima, A.B. Kulkarni, M. MacDougall, J. Bone Miner. Res. 1997, 12, 2040–2049. G. He, T. Dahl, A. Veis, A. George, Nat. Mater. 2003, 2, 552–558. L. Ye, Y. Mishina, D. Chen, H. Huang, S.L. Dallas, M.R. Dallas, P. Sivakumar, T. Kunieda, T.W. Tsutsui, A. Boskey, L.F. Bonewald, J.Q. Feng, J. Biol. Chem. 2005, 280, 6197–6203. K.N. Fedde, L. Blair, J. Silverstein, S.P. Coburn, L.M. Ryan, R.S. Weinstein, K. Waymire, S. Narisawa, J.L. Millan, G.R. MacGregor, M.P. Whyte, J. Bone Miner. Res. 1999, 14, 2015–2026. M.P. Whyte, Endocr. Rev. 1994, 15, 439–461. H. Fleisch, S. Bisaz, Nature 1962, 195, 911. R.A. Terkeltaub, Am. J. Physiol. Cell. Physiol. 2001, 281, C1–C11. R. Jemmerson, M.G. Low, Biochemistry 1987, 26, 5703–5709.

54 D.W. Moss, R.H. Eaton, J.K. Smith,

55

56

57

58

59

60

61 62

63 64 65

66

67

L.G. Whitby, Biochem. J. 1967, 102, 53–57. A.A. Rezende, J.M. Pizauro, P. Ciancaglini, F.A. Leone, Biochem. J. 1994, 301, 517–522. L. Hessle, K.A. Johnson, H.C. Anderson, S. Narisawa, A. Sali, J.W. Goding, R. Terkeltaub, J.L. Millan, Proc. Natl. Acad. Sci. USA 2002, 99, 9445–9449. R. Terkeltaub, M. Rosenbach, F. Fong, J. Goding, Arthritis Rheum. 1994, 37, 934–941. K. Johnson, A. Moffa, Y. Chen, K. Pritzker, J. Goding, R. Terkeltaub, J. Bone Miner. Res. 1999, 14, 883–892. A. Okawa, I. Nakamura, S. Goto, H. Moriya, Y. Nakamura, S. Ikegawa, Nat. Genet. 1998, 19, 271–273. F. Rutsch, N. Ruf, S. Vaingankar, M.R. Toliat, A. Suk, W. Hohne, G. Schauer, M. Lehmann, T. Roscioli, D. Schnabel, J.T. Epplen, A. Knisely, A. Superti-Furga, J. McGill, M. Filippone, A.R. Sinaiko, H. Vallance, B. Hinrichs, W. Smith, M. Ferre, R. Terkeltaub, P. Nurnberg, Nat. Genet. 2003, 34, 379–381. A.M. Ho, M.D. Johnson, D.M. Kingsley, Science 2000, 289, 265–270. P. Nurnberg, H. Thiele, D. Chandler, W. Hohne, M.L. Cunningham, H. Ritter, G. Leschik, K. Uhlmann, C. Mischung, K. Harrop, J. Goldblatt, Z.U. Borochowitz, D. Kotzot, F. Westermann, S. Mundlos, H.S. Braun, N. Laing, S. Tinschert, Nat. Genet. 2001, 28, 37–41. B. Bielesz, K. Klaushofer, R. Oberbauer, Bone 2004, 35, 1229–1239. A.J. Turner, K. Tanzawa, FASEB J. 1997, 11, 355–364. E.M. Eicher, J.L. Southard, C.R. Scriver, F.H. Glorieux, Proc. Natl. Acad. Sci. USA 1976, 73, 4667–4671. L. Beck, Y. Soumounou, J. Martel, G. Krishnamurthy, C. Gauthier, C.G. Goodyer, H.S. Tenenhouse, J. Clin. Invest. 1997, 99, 1200–1209. T.M. Strom, F. Francis, B. Lorenz, A. Boddrich, M.J. Econs, H. Lehrach, T. Meitinger, Hum. Mol. Genet. 1997, 6, 165–171.

References 68 H.S. Tenenhouse, Nephrol. Dial. 69 70 71

72

73 74

75

76

77 78

79 80

81

82

Transplant. 1999, 14, 333–341. P.S. Rowe, Crit. Rev. Oral Biol. Med. 2004, 15, 264–281. R.A. Meyer, Jr., M.H. Meyer, R.W. Gray, J. Bone Miner. Res. 1989, 4, 493–500. T. Nesbitt, T.M. Coffman, R. Griffiths, M.K. Drezner, J. Clin. Invest. 1992, 89, 1453–1459. H.S. Tenenhouse, A. Werner, J. Biber, S. Ma, J. Martel, S. Roy, H. Murer, J. Clin. Invest. 1994, 93, 671–676. H.S. Tenenhouse, L. Beck, Kidney Int. 1996, 49, 1027–1032. B. Ecarot, F.H. Glorieux, M. Desbarats, R. Travers, L. Labelle, J. Bone Miner. Res. 1992, 7, 215–220. Z.S. Xiao, M. Crenshaw, R. Guo, T. Nesbitt, M.K. Drezner, L.D. Quarles, Am. J. Physiol. 1998, 275, E700–E708. A. Gupta, H.S. Tenenhouse, H.M. Hoag, D. Wang, M.A. Khadeer, N. Namba, X. Feng, K.A. Hruska, Bone 2001, 29, 467–476. The ADHR Consortium, Nat. Genet. 2000, 26, 345–348. K.E. White, G. Carn, B. LorenzDepiereux, A. Benet-Pages, T.M. Strom, M.J. Econs, Kidney Int. 2001, 60, 2079–2086. N. Weidner, D. Santa Cruz, Cancer 1987, 59, 1442–1454. T. Shimada, S. Mizutani, T. Muto, T. Yoneya, R. Hino, S. Takeda, Y. Takeuchi, T. Fujita, S. Fukumoto, T. Yamashita, Proc. Natl. Acad. Sci. USA 2001, 98, 6500–6505. T. Shimada, I. Urakawa, Y. Yamazaki, H. Hasegawa, R. Hino, T. Yoneya, Y. Takeuchi, T. Fujita, S. Fukumoto, T. Yamashita, Biochem. Biophys. Res. Commun. 2004, 314, 409–414. K.B. Jonsson, R. Zahradnik, T. Larsson, K.E. White, T. Sugimoto, Y. Imanishi, T. Yamamoto, G.

83

84

85

86

87

88

89

90

91

92

Hampson, H. Koshiyama, O. Ljunggren, K. Oba, I.M. Yang, A. Miyauchi, M.J. Econs, J. Lavigne, H. Juppner, N. Engl. J. Med. 2003, 348, 1656–1663. Y. Yamazaki, R. Okazaki, M. Shibata, Y. Hasegawa, K. Satoh, T. Tajima, Y. Takeuchi, T. Fujita, K. Nakahara, T. Yamashita, S. Fukumoto, J. Clin. Endocrinol. Metab. 2002, 87, 4957– 4960. T.J. Weber, S. Liu, O.S. Indridason, L.D. Quarles, J. Bone Miner. Res. 2003, 18, 1227–1234. R. Guo, S. Liu, R.F. Spurney, L.D. Quarles, Am. J. Physiol. Endocrinol. Metab. 2001, 281, E837–E847. S. Liu, R. Guo, L.G. Simpson, Z.S. Xiao, C.E. Burnham, L.D. Quarles, J. Biol. Chem. 2003, 278, 37419– 37426. P.S. Rowe, P.A. de Zoysa, R. Dong, H.R. Wang, K.E. White, M.J. Econs, C.L. Oudet, Genomics 2000, 67, 54– 68. L. Argiro, M. Desbarats, F.H. Glorieux, B. Ecarot, Genomics 2001, 74, 342–351. R. Guo, P.S. Rowe, S. Liu, L.G. Simpson, Z.S. Xiao, L.D. Quarles, Biochem. Biophys. Res. Commun. 2002, 297, 38–45. S. Liu, T.A. Brown, J. Zhou, X.S. Xiao, H. Awad, F. Guilak, L.D. Quarles, J. Am. Soc. Nephrol. 2005, 16, 1645–1653. T. Shimada, M. Kakitani, Y. Yamazaki, H. Hasegawa, Y. Takeuchi, T. Fujita, S. Fukumoto, K. Tomizuka, T. Yamashita, J. Clin. Invest. 2004, 113, 561–568. S. Liu, J. Zhou, W. Tang, X. Jiang, D.W. Rowe, L.D. Quarles, Am. J. Physiol. Endocrinol. Metab. 2006, 291, E38–E49.

17

19

2 Bone Morphogenetic Proteins Walter Sebald, Joachim Nickel, Axel Seher, and Thomas D. Mu¨ller

Abstract

Bone morphogenetic proteins (BMPs) and growth and differentiation factors (GDFs) determine multiple processes in early embryonic development and organogenesis, including the formation of the bones and joints of the axial and appendicular skeleton. BMP-2 and BMP-7 (which is also known as OP-1) are approved as drugs and medical devices for the treatment of non-union fractures and for spinal fusion. Possible future indications for BMPs, GDFs or variants of these proteins include osteoporosis, osteoarthritis, fibrosis, parodontosis, and sinus lift. BMPs and GDFs are members of the TGF-b superfamily which signal into the cell by using two types of single-span membrane receptor chains that both have a cytosolic serine/threonine protein kinase domain. The extracellular ligand-binding domains are small, rich in disulfide bonds, and their fold is related to the three-finger toxins as, for example, are some of the conotoxins. The crystal structure of binary and ternary complexes between BMPs and the ectodomains of type I and type II receptors reveals the mechanism of receptor activation and the important determinants (hot spots) for binding specificity and affinity. Structure-based design of BMP and GDF variants yields proteins with new and potentially useful properties. Key words: bone morphogenetic protein (BMP), growth and differentiation factor (GDF), TGF-b superfamily, BMP receptor, X-ray structure, BMP signaling, bone formation and repair, mutants and variants.

2.1 Introduction

Orthopedic surgeons have long known that bone has an enormous capacity not only to heal fractures but also to regenerate defects up to a critical size. Furthermore, critical size defects that do not heal spontaneously can be repaired by Handbook of Biomineralization. Edited by M. Epple and E. Baüerlein Copyright 8 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ISBN: 978-3-527-31806-3

20

2 Bone Morphogenetic Proteins

implants of autologous or heterologous donor bone. Bone fragments that have been processed to different extents still result in the repair of critical size defects by processes of osteoconduction and de-novo osteoinduction. Most remarkably, bone fragments that have been demineralized and extracted for immunogenic proteins retain their osteoinductive properties. Mainly as a result of the investigations conducted by Marshal Urist [1], which commenced during the 1960s, it became clear that the osteoinductive properties of bone can be dissociated and extracted on a preparative scale from the collagenous bone matrix by chaotropic agents such as 6 M urea. This orthopedic research converged during the late 1980s with biochemical and recombinant DNA techniques. Starting with 100 kg of bovine bone, Wang et al. were able to isolate a few micrograms of protein that, after proteolytic digestions, yielded a few peptide sequences. After translation into DNA probes, genomic DNA and cDNAs of several BMP isoforms (BMP-2/-3/-4/-5/-6/-7/-8) could be obtained first from bovine and later from human sources [2, 3]. In some slightly earlier studies, Reddi, Sampath and colleagues succeeded in isolating pure OP-1 [4], which turned out to be the same protein as BMP-7. Pure natural OP-1, as well as recombinant BMP-2 to -7 produced in Chinese hamster ovary (CHO) cells possessed high osteoinductive activity when implanted with a carrier into a critical size defect, or into an ectopic site of an animal. It is clear today that single BMPs can induce in vivo the whole process of endochondrial or desmal bone formation, including vasculature and bone marrow. Recently, recombinant BMP-2 and OP-1/BMP-7 have been approved as drugs for orthopedic indications (non-union fractures, spinal fusion), and have acquired a considerable share of the bone reconstruction and tissue engineering market within a few years. BMP sequences are related to that of the transforming growth factor (TGF)-bs [5], the activins [6], and anti-Mu¨llerian hormone (AMH) [7], which had been identified prior to 1989. Together with the growth and differentiation factors (GDFs) [8] and glial-derived neurotrophic factor (GDNF) [9], these proteins now form the TGF-b superfamily which comprises more than 30 members. The TGFb-like proteins play pivotal roles during early and all later stages of embryonal development. In the adult organism, they regulate homeostasis and the repair of many tissues and organs. Despite these numerous functions of BMPs in many diverse tissues, the signaling machinery including ligands, receptors, and intracellular signaling proteins is remarkably conserved within this superfamily [10–13]. The role of BMPs in regenerative medicine [14], developmental biology [15], as well as their genetics in mouse mutants [16] and their role in the pathomechanism of human diseases [17], have been described in several excellent reviews. Recently an entire issue of Cytokine and Growth Factor Reviews [Volume 16(3)] has been devoted to BMPs, and today additional data on the structural basis of BMP signaling, on BMP regulation by modulator proteins, and on engineered BMP variants allow the subject to be discussed on a more molecular and chemical perspective with regard to the role(s) of BMPs, and what can be done with them.

2.2 What is a Bone Morphogenetic Protein?

2.2 What is a Bone Morphogenetic Protein?

The role of BMPs within the organism and their effects when provided externally are not the same. For example, whilst some BMPs are termed BMPs they have no bone-inducing capacity in vivo. BMP-1 is a procollagen proteinase [18], while BMP-3 is an inhibitor of bone formation [19]. Likewise, some members of the TGF-b superfamily are not called BMPs but rather GDFs, and they are (e.g., GDF-5) nevertheless supposed to be bone inducers [20]. The phylogenetic tree of the superfamily (Fig. 2.1A) shows that several subfamilies exist. In the following sections, the members of BMP-2/-4 subgroups will be referred to as BMP-2s, those of the BMP-5/-6/-7/-8 subgroups as BMP-7s, and those of the GDF-5/-6/-7 subgroups as GDF-5s. Some of the proteins have several designations, as they have been identified more than once, sometimes in different organisms. Thus, osteogenin 1 (OP-1) was described first by Reddi and co-workers [4], and independently as BMP-7 by Wozney and colleagues [2, 21]. GDF-5 was first described by Pohl and co-workers [22], but designated ‘‘cartilagederived morphogenetic protein 1’’ by another group [23]. The BMP-2s, BMP-7s and GDF-5s have been shown to induce ectopic bone formation and to induce alkaline phosphatase as a marker for osteoblastic differentiation in several cell lines. The sequences of BMP-2, OP-1 and GDF-5 are protected by patents which provided the basis for activities of three biotechnological companies in the orthopedic and dental markets. As detailed below, however, the proteins have distinct biochemical properties and receptor usage, and they are expressed at distinct times in specific tissues and cells. In addition to their bone-inducing capabilities, it is not yet clear how similar the function and activities of these proteins as therapeutics are within the organism. Members of the TGF-b superfamily are primarily synthesized as larger proproteins which initially dimerize, and are then cleaved at a RXXR site by a Furin-type protease. The C-terminal approximately 100 amino acid residues, as a dimer, represent the functional mature protein. Although the BMPs are normally homodimers, there is some indication that heterodimers between BMP-2 and BMP-6 or BMP-7 are significantly more active in vitro and in vivo [24]. The propiece as a dimer remains associated with some mature BMPs, without inhibiting their activity [25]. In contrast, the propiece of GDF-8 is a strong inhibitor of GDF-8 activity both in vitro and in vivo [26]. The BMPs and other members of the TGF-b superfamily have a very similar backbone fold and dimer assembly [27, 28]. The backbone forms a cystine knot, where the cysteine side chains of the signature CxGxC and CXC sequence elements constitute a ring, with a third disulfide bond threading through the opening of the ring. Further elements of the monomer fold are finger 1 and finger 2, each of which comprises a two-stranded b-sheet, a central a-helix and a prehelix loop, and finally an N-terminal peptide preceding the first cysteine of the knot (Fig. 2.1B,C). The whole monomer is usually compared to an open left hand.

21

22


Fig. 2.1 (A) Similarity tree of the human TGF-b superfamily, employing aligned sequences of the mature proteins (SwissProt; MultAlin). The average distance tree was constructed by means of the program ‘‘Jalview’’. The designations are BMP (bone morphogenetic protein), GDF (growth and differentiation factor), CDMP (cartilagederived morphogenetic protein), TGF-b (transforming growth factor beta), Act

(activin), Inh (inhibin), AMH/MIS (antiM€ ullerian hormone), GDNF (glial-derived neurotrophic protein), NRTN (neurturin), Mic (macrophage inhibitory cytokine). (B) Ribbon model of dimeric human BMP-2 [28]. Type I receptors bind to the ‘‘wrist epitope’’, and type II receptors to the ‘‘knuckle epitope’’ [53]. (C) Secondary structures and topology of the BMP-2 monomer.

The dimer is formed by mutually inserting the a-helix as the wrist region of one monomer into the concave open-hand region of the other monomer. An intermonomer disulfide bond stabilizes the dimer, and an axis of rotation runs between the two monomers. The BMPs are highly stable, and resistant to denaturation by chaotropic agents; this is most likely due to the extensive disulfide binding within and between the monomers. The BMP-2s and BMP-7s – but not the GDF-5s – are glycosylated at multiple sites. Glycosylation has only a minor effect on the solubility of the proteins, as

2.3 BMP Receptors are Composed of Diverse Type I and Type II Receptor Chains

non-glycosylated proteins – either naturally or after Escherichia coli expression – are insoluble under physiological conditions at concentrations higher than 0.1 to 0.25 mM. Glycosylated proteins, for example BMP-4 or BMP-7, remain soluble at two- to fourfold higher concentrations.

2.3 BMP Receptors are Composed of Diverse Type I and Type II Receptor Chains

In mammals, seven type I receptors and five type II receptors exist for TGF-b-like ligands (for reviews, see [10, 11, 29]). A structure-based alignment of the amino acid sequences of the extracellular domains [30] shows that the members of each type fall into one branch of an average distance sequence tree (Fig. 2.2). The established BMPs receptor chains BMPR-IA and BMPR-IB form a sub-branch, whereas ActR-I (Alk2) shows a relationship only to Alk1 (which is a TGF-b receptor). The dual-specificity receptors ActR-II and ActR-IIB which signal with both activines and BMPs form a sub-branch, whereas BMPR-II is only distantly related to all other type II chains. The ectodomains of the receptor chains are small and consist of only one domain of about 100 residues. They are rich in disulfide bonds, are connected to the membrane-spanning segment by a short linker, and they contain one to three putative N-glycosylation sites. The similarities among the amino acid sequences are very low. The pattern of the cysteines and disulfide bridges, however, is conserved and allows sequence alignment based on the established structures of ActR-II [31], TbR-II [32], BMPR-IA [30], and ActR-IIB [33]. The fold of the type I and type II ectodomains (Fig. 2.2) is characterized by one two-stranded and a three-stranded b-sheet, that are held together by five disulfide bonds. The type II ectodomains contain a third two-stranded sheet, whereas the

Fig. 2.2 Similarity tree of the human receptor chains for BMPs and other TGF-b-like proteins, employing structure-based sequence alignment of the ectodomains [30]. Left side: Ribbon models of the ectodomains of BMPR-IA [30] and ActR-II [31].

23

24


corresponding segment of the type I ectodomain is folded into a short a-helix. The ectodomains exhibit similarities with the cysteine-rich toxins, as for example conotoxins [34].

2.4 The Basic Signaling Mechanism is the Same for BMPs and other TGF-b-like Proteins

BMPs signal to cells by recruiting two types of single-span membrane receptors into a transphosphorylating complex (Fig. 2.3) (for reviews, see [35–37]. Both types of receptor chain have a cytosolic serine/threonine kinase domain. The type II kinase seems to be constitutively active and transphosphorylates a glycine/serine-rich region (GS box) of the type I chain at multiple sites, thereby activating its kinase. The type I receptors for BMPs (i.e., BMPR-IA, BMPR-IB and ActR-I) activate intracellular signaling Smad proteins which translocate into the nucleus to regulate the expression of target genes. In addition, non-Smad pathways can be activated [13]. The dimeric BMP ligand most likely binds two type I chains and two type II chains, and a hexameric complex consisting of

Fig. 2.3 Signaling BMP receptor complex consisting of type I and type II receptor chains. For further details, see text.

2.5 Biochemistry and Cell Biology of Receptor Specificity

BMP and two pairs of receptor chains must be assembled for efficient signal transduction [38]. BMPs share this conserved activation mechanism with the other members of the TGF-b superfamily. Several variations of the basic mechanism exist, however, as described below. For example, the BMP-2, BMP-7, and GDF-5 subgroups all use Smad 1, 5, or 8, whereas activines and TGF-bs use Smad 2 or 3 as the intracellular signaling protein. All receptor-restricted Smads (rSmads), with the exception of rSmad2, can directly bind to DNA, although the binding affinity is relatively low and cooperation with sequence-specific transcription factors is critical for the formation of a stable DNA-binding complex. For example, Smads 1 and 5 interact with bone-specific transcription factor Runx2 and activate the transcription of target genes [39]. About 500 target genes are estimated according to gene array data to be regulated by BMP signaling in the whole organism (see [10]). About 100 genes responsive to BMP signaling are found in C2C12 cell line upon osteoblastic differentiation [40].

2.5 Biochemistry and Cell Biology of Receptor Specificity

The basic signaling machinery and mechanism described above appears to be operating in all receptors for BMPs, and for the other TGF-b-like proteins. However, it is still not clear in all cases which BMPs or GDFs do functionally interact with which receptor chains, although a defined set of receptors has been identified as Smad1, 5, 8 activating and therefore as bona fide BMP receptors Pioneering experiments conducted a decade or so ago showed that BMPs can be chemically crosslinked in whole cells to the three type I receptors BMPR-IA, BMPR-IB and ActR-IA, and to the three type II receptors BMPR-II, ActR-II, and ActR-IIB [41–44]. The ActR-II and ActR-IIB receptors were originally identified as receptors for activin [45]. Later, these proteins were shown to exhibit dual specificity and to function also as BMP receptors [46]. ActR-I (Alk2) was first postulated to be an activin receptor, but later it was shown that ActR-I functions in BMP and not in activin signaling and activates Smads 1, 5, and 8 [46, 47]. Crosslinking experiments with transfected and non-transfected cell lines demonstrate that receptor preferences exist for BMPs from different subgroups, or even for members of the same subgroup. In the cell, BMP-2 and -4 use BMPRIA and possibly BMPR-IB as high-affinity receptors, but the role of ActR-I as a receptor for BMP-2s is uncertain [41]. Among the type II receptors, BMPR-II and ActR-II function with BMP-2s, but whether ActR-IIB is a receptor for BMP-2 or BMP-4 in vivo is unclear. The main GDF-5 receptor is BMPR-IB [48]. GDF-5 is much less efficient with BMPR-IA, and this is supported by the observation that, for example in C2C12 cells which contain BMPR-IA but are devoid of significant BMPR-IB levels,

25

26

2 Bone Morphogenetic Proteins Table 2.1 Dissociation constants (KD ) for the 1:1 interaction of receptor ectodomains with biosensor-immobilized BMPs/GDFs (Biacore analysis).

Receptor

BMPR-IA BMPR-IB ActR-I BMPR-II ActR-II ActR-IIB

KD [mM] BMP-2

BMP-7

GDF-5

0.015* 0.095* 420 54 14 8.5

10 1.1* 55 32 0.43 2.7

3.3* 0.3* – 60 22 4.7

* Dissociation

constants were calculated from the kinetic constants as KD ¼ koff =kon . All other KD values were evaluated form the dosedependence of equilibrium binding.

GDF-5 is inactive. GDF-5 signaling is enabled by transfection with BMPR-IB; GDF-5 can also use BMPR-II or ActR-II as type II receptors. BMP-7 and BMP-6 can signal efficiently in cells that are devoid of BMPR-IB or of both BMPR-IA and BMPR-IB, whereas BMP-4 cannot, thereby indicating that ActR-I is a functional receptor for these ligands. Interaction with BMPR-II and ActR-II seems to be possible. Detailed investigations established that large differences in binding affinities exist for different ligand/receptor constellations. The apparent dissociation constants, KD , for the interaction of receptor ectodomains with immobilized ligands as determined by biosensor analysis (Biacore) are listed in Table 2.1. BMP-2 binds preferentially to BMPR-IA, and six- to 10-fold less tightly to BMPR-IB. The affinity for ActR-I and for all type II receptors is over 100-times lower. GDF-5 as the representative of another subgroup binds with intermediate affinity to BMPR-IB, and 10-fold less tightly to BMPR-IA. All type II ectodomains bind to GDF-5 with very low affinity. BMP-7, as a member of a third subgroup, exhibits no clear highaffinity interaction with any of these receptors. All three type II chains bind with about micromolar affinity, and a medium affinity is seen with BMPR-IB. Remarkably, BMP-6 and -7, as members of the same subgroup, show differences in binding affinity for some chains, such as BMPR-IA, BMPR-IB, or ActR-II. The in-situ crosslinking and the ectodomain binding data for BMP-2 can be reconciled on the basis of the following two-step mechanism. The activation of these receptors is a sequential process: first, the ligand binds to its high-affinity receptor chain – that is, BMPR-IA (or if necessary BMPR-IB) for BMP-2. Second, in the membrane the low-affinity type II chain is recruited into the signaling complex. Collisions – and therefore also on-rates – occur more frequently in the membrane

2.6 Structural Basis for Specificity and Affinity of BMP Receptor Interaction

than in solution, and therefore solute KD s in excess of 1 mM can be kinetically competent for productive transactivation by reduction of dimensionality. One essential feature of the sequential two-step mechanism is the high-affinity binding of solute BMP-2. However, the affinities of GDF-5, BMP-6 or -7 for any of the receptor chains being discussed here appear to be much too low for the efficient binding of a solute ligand. In particular, the extreme low affinity of ActR-I for BMP-6 and BMP-7 certainly requires some high-affinity interaction with solute ligand. The in-vitro data in Table 2.1 were acquired by measuring 1:1 interactions between ectodomains and the ligands. However, it remains uncertain as to how the affinities change when the dimeric ligand binds to the membrane receptors, which exist to a large percentage in oligomeric form (avidity effects) [49]. As another possibility, accessory or modulatory proteins might cooperate with type II receptors to generate high-affinity binding. (A similar situation exists in the TGFb receptors, where betaglycan or endoglin converts the low-affinity interaction between TGF-b2 and TbRII into a high-affinity signaling-competent interaction.)

2.6 Structural Basis for Specificity and Affinity of BMP Receptor Interaction

Binary and ternary complexes between BMP-2 or BMP-7 and ectodomains of BMPR-IA, ActR-II and ActR-IIB, have been crystallized and their structures elucidated [30, 50–52]. These data yield important insights into the assembly mechanism of, and molecular recognition in, BMP receptor complexes. The structures also provide the basis for the design of BMP-2 variants with new and potentially useful properties. BMPs are homodimeric and consequently can bind two type I and two type II receptors, thereby forming a ternary complex with a pseudohexameric 1:2:2 composition. Because BMP-2 is a dimer of two identical monomers, the complete complex consists of six polypeptides chains (Fig. 2.4A). BMP-2, with its two hand-like monomers, binds its BMPR-IA receptor at the so-called ‘‘wrist epitope’’ [30, 53] (Fig. 2.4B). This epitope is composed of segments from two monomers; one monomer contributes the central a-helix plus its preceding loop, and the other contributes the concave side of the two twostranded b-sheets. The resultant surface forms a hydrophobic patch with a deep hole on one side. The interface with BMPR-IA shows that 10 hydrogen bridges exist between receptor and ligand. Mutational analysis of contact residues [54] revealed that the two central hydrogen bridges involving main-chain NH- and COgroups are important binding determinants. Interestingly, these polar bonds originate in the pre-helix loop of BMP-2, and one of them ends in the Gln86 side chain of the receptor located in the short a-helix. In addition to these polar bonds, many hydrophobic interactions are found at the interface. The most conspicuous hydrophobic contact exists between the hole of BMP-2 and the Phe85 side chain protruding from the a-helix of the receptor (‘‘knob-into-hole’’ motif ). Thus, the

27

28


Fig. 2.4 (A) Ribbon model of the crystal structure of the ternary complex consisting of BMP-2 and two ectodomains of BMPR-IA and two ectodomains of ActR-IIB [52]. (B) Open-book view of the interface between BMPR-IA and BMP-2 wrist epitope. (C) Interface between ActR-IIB and BMP-2 knuckle epitope. Residues are color-coded: gray ¼ aliphatic; green ¼ polar; red ¼ acidic; blue ¼ basic side chains.

2.7 What We Can Do with BMPs: The Engineering of BMP-2 and GDF-5 Variants

a-helix of the type I receptor carries the most important polar and hydrophobic determinants for binding. Until now, the BMP-2 wrist epitope for BMPR-IA binding is the only example of type I receptor interaction. However, it is of interest to note that large knob-like side chains exist in all type I receptors of the TGF-b superfamily, with the exception of Alk1 at a position equivalent to BMPR-IA Phe85. Furthermore, deep holes are seen in the structures of all known TGF-b-like proteins, with the possible exception of BMP-9 [55]. It is therefore conceivable that all of these type I receptors bind their ligands at a site equivalent to the BMP-2 wrist epitope. BMP-2 binds the type II ActR-II and ActR-IIB receptors at the so-called ‘‘knuckle’’ epitope [51, 52] (Fig. 2.4C). Activin A, another TGF-b-like protein, binds ActR-IIB at the same site [33, 56]. Because BMP-7 also uses this site for ActR-II interaction [50], it seems safe to conclude that the knuckle epitope is the common site for BMP and activin interaction with the type II receptors ActR-II and ActR-IIB. Certain mutations in this epitope enhance BMPR-II binding to BMP-2 [53]; thus, all type II receptors are likely to bind to this epitope. (The situation is different in TGF-b interaction with TbRII, where the binding epitope of TGF-b is located at the finger tips; that is, at the terminal loop regions of the b-sheets [57].) The knuckle epitope occurs at the convex side of the BMP finger region, and is formed from residues of one monomer only. The contact is predominantly hydrophobic, with side chains of aromatic residues at the receptor side and of aliphatic residues at the ligand side. A conserved hydrogen bond at the core of the knuckle epitope contact provides an enlightening example of how high- and low-affinity interactions are generated in the various receptor systems [51, 52]. When this hydrogen bond is mutationally disrupted in ActR-IIB, high-affinity binding to activin A is lost, with the affinity being reduced to the low levels observed in BMP-2 interaction. In contrast, disruption of the corresponding interaction of BMP-2 has only a marginal effect on BMP binding. The geometry of the hydrogen bond is similar in both circumstances, but a conspicuous difference exists for the two side chains flanking the bond at the periphery. A Lys/Asp pair provides perfect sealing from the solute in the ActA/ActR-IIB contact, whereas a Leu/Asn pair seems to allow some solute access. Transferring the ActA side chains to BMP-2 generates high-affinity binding to ActR-IIB. X-ray structure investigations reveal, indeed, that the Lys/Asp side chains of mutated BMP-2 are immobilized by forming an ion pair, and that the hydrophobic part of these residues provides a perfectly sealed contact [52].


The N-terminal segment of BMP-2 carries a heparin-binding epitope [42, 58]; these segments are flexible and present twice in the dimeric BMP, and therefore

29

30


Fig. 2.5 Direct reconstitution of the mandible bone of a mini-pig. (A) A critical-size 5-cm defect in the mandible was treated with carrier material plus recombinant BMP-2 [67]. Full regeneration of the mandible with a mechanically stable bone is visible in the

X-rays taken after 8 weeks. (B) The control defect treated with carrier alone formed a pseudarthrose, i.e. the defect was filled with connective tissue. (Reproduced from Ref. [67], with permission of Springer Sciences and Business Media.)

they can provide tight links to glycosaminoglycans present in the extracellular matrix (ECM) or the plasma membrane. This interaction likely localizes BMP activity for autocrine or paracrine functions. A BMP-2 variant in which the heparin-binding epitope has been removed [58] showed reduced activity during ectopic bone formation [59]. In contrast, BMP-2 variants with a duplicated heparin-binding epitope showed stronger binding to heparin in vitro and a more efficient bone formation in vivo [59]. The functional reconstitution of large bone defects can be accomplished with recombinant BMP2, as shown in Figure 2.5. Another type of BMP-2 variant with mutations in the knuckle epitope shows a drastic loss in type II receptor binding affinity [53]. The A34D variant has retained very low biological activity, but inhibits the activity of normal BMP-2; this variant is thus an antagonist that competes with normal BMP2 for BMPR-IA receptor binding. Remarkably, the inhibitory activity (IC50 ) of the antagonist is similar to the activity (EC50 ) of normal BMP-2, which indicates that the type II receptor contributes very little – if at all – to the receptor affinity in whole cells. This variant represents a receptor antagonist, whereas natural BMP inhibitors (e.g., Noggin [60], which has been used to inhibit BMP activity in vivo (e.g., [61]), bind directly to the ligand. The wrist epitope of BMP-2, which determines type I receptor binding, can be inactivated by a substitution of Leu51 by proline [54]. The resultant L51P variant


Fig. 2.6 (A) Familial symphalangism caused by a gain-of-function mutation in GDF-5 [66]. Joints are replaced by bone in finger V and defective in finger IV (see arrows). The gene exhibits a mutation R438L located in the wrist epitope of GDF-5 (R57L in the mature

protein). The mutant GDF-5 has a severalfold increased affinity for the BMPR-IA receptor. (B) A similar phenotype is produced by heterozygous mutations in the Noggin gene. (Reproduced from Ref. [66], with kind permission.)

is receptor-dead, as it binds only very weakly to the type I receptors. The defect produced by this substitution is local and allows the remainder of the BMP-2 surface to interact with a variety of other proteins, including many BMP modulator proteins. Indeed, the L51P variant releases the inhibition by Noggin and other BMP inhibitors for BMP-2-dependent functions. The L51P variant could therefore be useful during conditions such as fractures [62], osteoporosis [63], and osteoarthritis [64], in which BMP activity might to be limited by modulator proteins as Noggin, CTGF, CHL2, and others. Finally, a GDF-5 variant has been created that exhibits altered receptor specificity [65]. Normal GDF-5 binds preferentially to the BMPR-IB receptor [48], and only with a much lower affinity to BMPR-IA. A gain-of-function variant of GDF5 could be created by substituting Arg57 by alanine, with the resultant R57A variant binding with identical high affinity to both BMPR-IB and BMPR-IA. It might be of interest to determine if this change in receptor specificity will influence the in-vivo activity of GDF-5, for example during bone induction. Remarkably, a similar gain-of-function mutation R57L has been identified in a form of familiar symphalangism characterized by missing joints in the distal phalanges (Fig. 2.6) [66]. Here, the phenotype is similar to that produced by inactivating mutations in the noggin gene. Inactivating the inhibitor Noggin will also lead to increased BMP signaling, similar to that seen with the gain-of-function GDF-5 mutant. The crystal structures of wild-type and variants of BMPs and GDFs show that the backbone fold and dimer assembly of these proteins is largely resilient to the type of side chains in the receptor-binding wrist and knuckle epitopes. Even proline substitutions are tolerated. The dimeric cystine-knot scaffold seems to stabilize these proteins so that specificities and affinities for diverse receptors can be created by exposing diverse side-chain patterns on an invariant backbone.

31

32


References 1 M.R. Urist, Science 1965, 150, 893. 2 E.A. Wang, V. Rosen, P. Cordes, R.M.

3

4

5

6

7

8 9

10 11

12 13 14 15 16 17 18

19

Hewick, M.J. Kriz, D.P. Luxenberg, B.S. Sibley, J.M. Wozney, Proc. Natl. Acad. Sci. USA 1988, 85, 9484. J.M. Wozney, V. Rosen, A.J. Celeste, L.M. Mitsock, M.J. Whitters, R.W. Kriz, R.M. Hewick, E.A. Wang, Science 1988, 242, 1528. F.P. Luyten, N.S. Cunningham, S. Ma, N. Muthukumaran, R.G. Hammonds, W.B. Nevins, W.I. Woods, A.H. Reddi, J. Biol. Chem. 1989, 264, 13377. R. Derynck, J.A. Jarrett, E.Y. Chen, D.H. Eaton, J.R. Bell, R.K. Assoian, A.B. Roberts, M.B. Sporn, D.V. Goeddel, Nature 1985, 316, 701. A.J. Mason, J.S. Hayflick, N. Ling, F. Esch, N. Ueno, S.Y. Ying, R. Guillemin, H. Niall, P.H. Seeburg, Nature 1985, 318, 659. R.L. Cate, R.J. Mattaliano, C. Hession, R. Tizard, N.M. Farber, A. Cheung, E.G. Ninfa, A.Z. Frey, D.J. Gash, E.P. Chow, et al., Cell 1986, 45, 685. S.J. Lee, Mol. Endocrinol. 1990, 4, 1034. L.F. Lin, D.H. Doherty, J.D. Lile, S. Bektesh, F. Collins, Science 1993, 260, 1130. Y. Shi, J. Massague, Cell 2003, 113, 685. S. Souchelnytskyi, A. Moustakas, C.H. Heldin, Trends Cell Biol. 2002, 12, 304. P. ten Dijke, C.S. Hill, Trends Biochem. Sci. 2004, 29, 265. A. Moustakas, C.H. Heldin, J. Cell Sci. 2005, 118, 3573. A.H. Reddi, Nat. Biotechnol. 1998, 16, 247. B.L. Hogan, Curr. Opin. Genet. Dev. 1996, 6, 432. G.Q. Zhao, Genesis 2003, 35, 43. X. Cao, D. Chen, Gene 2005, 357, 1. S.W. Li, A.L. Sieron, A. Fertala, Y. Hojima, W.V. Arnold, D.J. Prockop, Proc. Natl. Acad. Sci. USA 1996, 93, 5127. A. Daluiski, T. Engstrand, M.E. Bahamonde, L.W. Gamer, E. Agius,

20

21

22

23

24

25

26 27

28 29

30 31 32

33 34 35 36 37 38

S.L. Stevenson, K. Cox, V. Rosen, K.M. Lyons, Nat. Genet. 2001, 27, 84. R.C. Spiro, L. Liu, M.A. Heidaran, A.Y. Thompson, C.K. Ng, J. Pohl, J.W. Poser, Biochem. Soc. Trans. 2000, 28, 362. A.J. Celeste, J.A. Iannazzi, R.C. Taylor, R.M. Hewick, V. Rosen, E.A. Wang, J.M. Wozney, Proc. Natl. Acad. Sci. USA 1990, 87, 9843. G. Hotten, H. Neidhardt, B. Jacobowsky, J. Pohl, Biochem. Biophys. Res. Commun. 1994, 204, 646. S.C. Chang, B. Hoang, J.T. Thomas, S. Vukicevic, F.P. Luyten, N.J. Ryba, C.A. Kozak, A.H. Reddi, M. Moos, Jr., J. Biol. Chem. 1994, 269, 28227. D.I. Israel, J. Nove, K.M. Kerns, R.J. Kaufman, V. Rosen, K.A. Cox, J.M. Wozney, Growth Factors 1996, 13, 291. D.I. Israel, J. Nove, K.M. Kerns, I.K. Moutsatsos, R.J. Kaufman, Growth Factors 1992, 7, 139. S.J. Lee, A.C. McPherron, Proc. Natl. Acad. Sci. USA 2001, 98, 9306. P.R. Mittl, J.P. Priestle, D.A. Cox, G. McMaster, N. Cerletti, M.G. Grutter, Protein Sci. 1996, 5, 1261. C. Scheufler, W. Sebald, M. Hulsmeyer, J. Mol. Biol. 1999, 287, 103. P. ten Dijke, J. Fu, P. Schaap, B.A. Roelen, J. Bone Joint Surg. Am. 2003, 85-A (Suppl. 3), 34. T. Kirsch, W. Sebald, M.K. Dreyer, Nat. Struct. Biol. 2000, 7, 492. J. Greenwald, W.H. Fischer, W.W. Vale, S. Choe, Nat. Struct. Biol. 1999, 6, 18. C.C. Boesen, S. Radaev, S.A. Motyka, A. Patamawenu, P.D. Sun, Structure (Camb.) 2002, 10, 913. T.B. Thompson, T.K. Woodruff, T.S. Jardetzky, EMBO J. 2003, 22, 1555. H. Terlau, B.M. Olivera, Physiol. Rev. 2004, 84, 41. J. Massague, Annu. Rev. Biochem. 1998, 67, 753. C.H. Heldin, K. Miyazono, P. ten Dijke, Nature 1997, 390, 465. P. ten Dijke, K. Miyazono, C.H. Heldin, Curr. Opin. Cell Biol. 1996, 8, 139. P. Knaus, W. Sebald, Biol. Chem. 2001, 382, 1189.

References 39 K.S. Lee, H.J. Kim, Q.L. Li, X.Z. Chi,

40

41

42

43 44

45 46

47

48

49

50

51

52

53 54

C. Ueta, T. Komori, J.M. Wozney, E.G. Kim, J.Y. Choi, H.M. Ryoo, S.C. Bae, Mol. Cell. Biol. 2000, 20, 8783. O. Korchynskyi, K.J. Dechering, A.M. Sijbers, W. Olijve, P. ten Dijke, J. Bone Miner. Res. 2003, 18, 1177. P. ten Dijke, H. Yamashita, T.K. Sampath, A.H. Reddi, M. Estevez, D.L. Riddle, H. Ichijo, C.H. Heldin, K. Miyazono, J. Biol. Chem. 1994, 269, 16985. B.B. Koenig, J.S. Cook, D.H. Wolsing, J. Ting, J.P. Tiesman, P.E. Correa, C.A. Olson, A.L. Pecquet, F. Ventura, R.A. Grant, et al., Mol. Cell. Biol. 1994, 14, 5961. F. Liu, F. Ventura, J. Doody, J. Massague, Mol. Cell. Biol. 1995, 15, 3479. B.L. Rosenzweig, T. Imamura, T. Okadome, G.N. Cox, H. Yamashita, P. ten Dijke, C.H. Heldin, K. Miyazono, Proc. Natl. Acad. Sci. USA 1995, 92, 7632. L.S. Mathews, W.W. Vale, J. Biol. Chem. 1993, 268, 19013. H. Yamashita, P. Ten Dijke, C.H. Heldin, K. Miyazono, Bone 1996, 19, 569. M. Macias-Silva, P.A. Hoodless, S.J. Tang, M. Buchwald, J.L. Wrana, J. Biol. Chem. 1998, 273, 25628. H. Nishitoh, H. Ichijo, M. Kimura, T. Matsumoto, F. Makishima, A. Yamaguchi, H. Yamashita, S. Enomoto, K. Miyazono, J. Biol. Chem. 1996, 271, 21345. L. Gilboa, A. Nohe, T. Geissendorfer, W. Sebald, Y.I. Henis, P. Knaus, Mol. Biol. Cell 2000, 11, 1023. J. Greenwald, J. Groppe, P. Gray, E. Wiater, W. Kwiatkowski, W. Vale, S. Choe, Mol. Cell 2003, 11, 605. G.P. Allendorph, W.W. Vale, S. Choe, Proc. Natl. Acad. Sci. USA 2006, 103, 7643. D.Weber,A.Kotzsch,J. Nickel,S.Harth, A. Seher, K. Mueller, W. Sebald, T.D. Mueller, BMC Struct. Biol. 2007, 7, 6. T. Kirsch, J. Nickel, W. Sebald, EMBO J. 2000, 19, 3314. S. Keller, J. Nickel, J.L. Zhang, W. Sebald, T.D. Mueller, Nat. Struct. Mol. Biol. 2004, 11, 481.

55 M.A. Brown, Q. Zhao, K.A. Baker, C.

56

57

58 59

60

61

62

63

64

65

66

67

Naik, C. Chen, L. Pukac, M. Singh, T. Tsareva, Y. Parice, A. Mahoney, V. Roschke, I. Sanyal, S. Choe, J. Biol. Chem. 2005, 280, 25111–25118. J. Greenwald, M.E. Vega, G.P. Allendorph, W.H. Fischer, W. Vale, S. Choe, Mol. Cell 2004, 15, 485. P.J. Hart, S. Deep, A.B. Taylor, Z. Shu, C.S. Hinck, A.P. Hinck, Nat. Struct. Biol. 2002, 9, 203. R. Ruppert, E. Hoffmann, W. Sebald, Eur. J. Biochem. 1996, 237, 295. K.K. Wurzler, J. Emmert, F. Eichelsbacher, N.R. Kubler, W. Sebald, J.F. Reuther, Mund. Kiefer. Gesichtschir. 2004, 8, 83. J. Groppe, J. Greenwald, E. Wiater, J. Rodriguez-Leon, A.N. Economides, W. Kwiatkowski, M. Affolter, W.W. Vale, J.C. Belmonte, S. Choe, Nature 2002, 420, 636. E. Canalis, A.N. Economides, E. Gazzerro, Endocr. Rev. 2003, 24, 218. Y. Yoshimura, S. Nomura, S. Kawasaki, T. Tsutsumimoto, T. Shimizu, K. Takaoka, J. Bone Miner. Res. 2001, 16, 876. G. Turgeman, Y. Zilberman, S. Zhou, P. Kelly, I.K. Moutsatsos, Y.P. Kharode, L.E. Borella, F.J. Bex, B.S. Komm, P.V. Bodine, D. Gazit, J. Cell Biochem. 2002, 86, 461. N. Nakayama, C.Y. Han, L. Cam, J.I. Lee, J. Pretorius, S. Fisher, R. Rosenfeld, S. Scully, R. Nishinakamura, D. Duryea, G. Van, B. Bolon, T. Yokota, K. Zhang, Development 2004, 131, 229. J. Nickel, A. Kotzsch, W. Sebald, T.D. Mueller, J. Mol. Biol. 2005, 349, 933–947. P. Seemann, R. Schwappacher, K.W. Kjaer, D. Krakow, K. Lehmann, K. Dawson, S. Stricker, J. Pohl, F. Ploger, E. Staub, J. Nickel, W. Sebald, P. Knaus, S. Mundlos, J. Clin. Invest. 2005, 115, 2373. K.K. Wurzler, M. Heisterkamp, H. Bohm, N.R. Kubler, W. Sebald, J.F. Reuther, Mund. Kiefer. Gesichtschir. 2004, 8, 75.

33

35

3 Biomechanics of Bones: Modeling and Computation of Bone Remodeling Udo Nackenhorst

Abstract

Bones are living organs that have the ability to adapt themselves to their mechanical demands. This phenomenon is of major importance in endoprosthetics. Following an artificial implant, the bone is stressed in a non-physiological manner, and this causes bone remodeling. Computational methods are available to predict this behavior, which in turn allows the optimization of prosthesis design such that the surgeon can identify the best available implant for an individual patient’s condition. However, many uncertainties are encountered when quantifying the mechanical loading conditions and the overall mechanical properties of bone tissue. The concept of statically equivalent loads is stated, where the boundary conditions are computed by an inverse simulation from computed tomography data. The mechanical properties of cortical bone are obtained using a micromechanical approach, with several stages of homogenization. Moreover, the process of mechanotransduction may be simulated by using this multi-scale approach. Key words: stress-adaptive bone remodeling, finite element techniques, hip– joint endoprosthetics, multi-scale methods.

3.1 Introduction

Bone remodeling describes the mechanically driven process of changes in bone constitution with regards to their geometry and internal architecture. The most famous citation on this phenomenon is by Wolff [1], who stated that the structure of bones can be determined by mathematical rules, depending on their mechanical demands. Although Wolff did not propose any written formulae, his statement has been underlined by several investigations, whilst in the past mathematical models have also been derived. Handbook of Biomineralization. Edited by M. Epple and E. Baüerlein Copyright 8 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ISBN: 978-3-527-31806-3

36

3 Biomechanics of Bones: Modeling and Computation of Bone Remodeling

Inspired by this simple idea, somewhat phenomenological mathematical models and related computational techniques have been developed, enforced by the exponentially increasing availability of computational recourses, and beginning during the 1980s. The phenomenology relates to the growth mechanism, which is assumed to be driven by a rather simple mechanical stimulus. The state of the art of these phenomenological approaches will be reviewed in the first section of this chapter, where it will be shown that today such models permit the conduct of qualitative studies on bone remodeling caused by changes in stress conditions following endoprosthesis implant. Later, fully three-dimensional (3-D) analyses are performed, taking individual environmental conditions into account. For example, patient-specific geometry as well as equivalent muscle forces and joint loads can be derived from computed tomography (CT) data. These computational techniques may help in identifying optimized and biomechanically more compatible designs for prostheses, as well as determining the best implant and treatment procedure for an individual patient’s condition. These approaches will be of qualitative nature, mainly because in-vivo validation is not possible. A major question here relates to the mechanism of mechanical stimulus, which has been the subject of much controversy in reports [2, 3]. This process – which is referred to as called ‘‘mechanosensation’’ and ‘‘mechanotransduction’’ [4] – is driven by the bone cells, the so-called osteocytes that are embedded between the dense bone and connected by a network of numerous processes. Rather simple experiments with cell cultures have been conducted to investigate the nature of mechanosensation [5]. Recently, computations on the behavior of bone cells have been reported, on a cellular scale and with detailed modeling of the cytoskeleton and nucleus, down to a scale where the response of proteins due to mechanical forces has been studied [6, 7]. However, in order to obtain the complete picture, the cells must be studied in their natural environment. Because of the hierarchical architecture of bone (e.g., see [8]), a multi-scale analysis is necessary for these investigations. A first approach on such as computational multi-scale analysis is outlined in Section 3.2, and takes into account the osteonal structure of cortical bone, and the laminar architecture of the osteons with anisotropic material properties obtained on a third smaller length-scale, depending on the grade of mineralization. Rather simple cell models have been modeled between the lamellae for sensing strains and signaling the ongoing mineralization and growth of existing osteons, as well as for the creation of new osteons which are dependent upon mechanical demand.

3.2 The Biomechanical Equilibrium Approach

For single-scale macroscopic investigations of stress-adaptive bone remodeling, a continuum approach is suitable. In this way, the microstructure is smeared and expressed by an averaged bone mass density distribution, for cortical as well as spongious bone, which is related to constitutive equations (cf. [9–11]). Because


remodeling phenomena appear to function on long time scales in relation to an individual motion, a quasi-static and isothermal treatment is justified. This approach is referred to as the theory of ‘‘biomechanical equilibrium’’, because the target is a mechanical equilibrium state with the subsidiary condition that no change of bone mass appears anywhere. Based on this assumption, the problem can be defined mathematically as: (a) the mechanical equilibrium conditions of a continuous solid body; (b) a constitutive law which incorporates the remodeling phenomena; and (c) boundary conditions. The mechanical equilibrium (a) is described by elliptic partial differential equations, for which the finite element method (FEM) has been proven to provide efficiently approximate solutions. Problems on numerical stability have been discussed in the literature – and solved – with regards to the simulation of bone remodeling during its early stages (e.g., [12]). The constitutive law (b) for the biomechanical problem consists of several ingredients, which include the basic stress–strain relationships, a constitutive relationship between the mechanical properties and the bone mass density distribution, and an evolutionary rule for the mechanically driven local rate of change of bone mass density. With regards to the stress–strain relationship, linear elasticity is amicably accepted, where macroscopic anisotropic mechanical behavior might be taken into account [10, 13, 14]. With regards to the second aspect – namely the relationship between elastic coefficients and the scalar valued bone mass density – there is no consensus in literature to date. Here, often-cited studies include the empirical findings after Carter and Hayes [15], which relate Young’s modulus and apparent bone mass density by a power law. Other authors prefer piecewise power laws with fractional exponents which represent fits to measurements [9, 16]. By a statistical analysis based on much experimental data reported previously in the literature, Rice et al. [17] have shown that the quadratic term in a polynomial approximation behaves in dominant fashion, which means that Young’s modulus is proportional to the square of bone mass density. This latter finding has been underlined by theoretical investigations within a framework of consistent continuum theory of materials reported elsewhere [14]. Based on the entropy principle, it is argued that the relationship between Young’s modulus E and bone mass density r is of the form E @ r 2 . The difference in physical dimensions is treated by a dimensional constant, to be adjusted from experimental results. The isotropic approach sketched here, is easily extended to a more sophisticated theory, where anisotropic mechanical behavior is taken into account [14]. The final constituent of the continuum constitutive model is on the formulation of adequate rules for the stress-driven evolution of bone mass density. Rather heuristically, suggestions have been made where often strong non-linear, piecewise continuous evolution rules have been introduced [9, 16]. However, this subject can also be treated within the consistent framework of continuum theory of

37

38


materials, where it can be classified as locating between continuum damage mechanics, because of the explicit changes in elastic properties as a result of mechanical treatment (e.g., [10]), and visco-elasticity because of the explicit time dependency of the processes (see [18]). Within this theoretical framework, the evolutionary rule is derived from the entropy-principle, where the mechanical stimulus function is obtained as conjugate thermodynamic force which drives bone growth [18]. A remaining challenge here is to describe the boundary conditions (c) with regards to the muscle forces and joint loads. Despite the fact, that much excellent data have been based on measured hip-joint loads for specific activities [19] and related muscle force collectives have been computed [20], these data cannot be used directly for the remodeling simulation because they reflect short-term action only. The time scale for bone remodeling processes extends to month and years, and such an argument has already has been stressed in order to justify the quasistatic equilibrium approach. Consequently, a statically equivalent load set must be described which, in an ideal case, reflects the individual conditions. Related computational strategies have now been under development for more than 10 years (see [21]), and recently have been made available for practice on 3-D bone remodeling simulations [22]. When applying these approaches, which are based on measured bone mass density distributions (e.g., CT data), the statically equivalent muscle forces and joint loads are computed by solving the inverse problem, which is defined mathematically as follows: Find the muscle-forces and joint loads, such that a given bone mass density distribution is obtained in the biomechanical equilibrium state. A combination of genetic and gradient-type algorithms have been developed in order to compute the statically equivalent load sets for this ill-posed problem. A result of this schema is depicted in Figure 3.1. The measured bone mass densities for some horizontal slices are shown at the left of the figure. These results are mapped onto the finite element discretization of the femoral bone (as depicted in the center of the figure). With these data, the statically equivalent load sets are computed by solving the inverse problem. Finally, a straightforward iteration is performed to generate an equilibrated model. The resulting bone mass density distribution obtained for this example, taking into account five muscle forces, is depicted at the right of the figure. The correlation with the originally measured data was quite good, the differences being due mainly to the imaging based on the finite element discretization and the pixel-based CTimages. With regards to the sensitivity of the results on the finite element discretization, it has been shown in numerical experiments [22] that rather rough models lead to qualitatively good results. For example, in the presentation shown in Figure 3.1, fewer than 4000 linear tetrahedral elements were used, a fact which also underlines the robust nature of the computational approach. One point should be made with regards to the relevant muscle groups that have been taken into account. During the early investigations, it was shown that the joint load and subsumption of the abductor muscles acting at the trochanter major (a simple equilibrium of a statically determined, one-leg stand condition)


Fig. 3.1 Mapping of computed tomography (CT) data (left) onto the finite element mesh (middle) and reverse-computed bone mass density distribution (right) for the biomechanical equilibrium state.

leads to qualitatively good results [23]. The optimization techniques for solving the inverse problem provide clear hints on the relevance of different muscle groups in the bone remodeling simulation. For the example depicted in Figure 3.1, the statically equivalent load set listed in Table 3.1 has been computed. For comparison, maximal forces obtained for a walking sequence are listed. From these results it is clearly seen that, in addition to the joint force, the gluteus group provides the most dominant influence, followed by the psoas major and the

Table 3.1 Computed statically equivalent load sets in comparison with

maximal forces. (Modified from Ref. [20].) Indication

Statically equivalent forces [N]

Maximal forces during walking [N]

Joint force Gluteus minimus Gluteus medius Gluteus maximus Psoas major, illiacus Vastus lateralis Vastus intermedialis Biceps femoris, caput breve Vastus medialis Adductor longus

1371 589 272 223 222 139 58 21 7 4

2190 284 306 91 174 228 63 92 9 7

39

40


Fig. 3.2 Bone remodeling caused by a standard stem prosthesis. Compared to the immediate postoperative state (left), a clear loss of bone mass is observed in the cortical bone surrounding the stem in the biomechanical equilibrium state (right).

vastus group. In comparison with the peak forces reported for a walking sequence, the values differ significantly, although this is not a surprising result as the statically equivalent load set reflects an average long-term median. However, the tendency in the ordering scheme seems to be the same. From these results it can be concluded that at least five muscle groups must be taken into account for a sufficient bone remodeling simulation (cf. [24]). With these careful preparations – which are based on a rather phenomenological but nonetheless physically consistent approach – tools are available to study bone remodeling caused by medical treatment, for which total hip-joint arthroplasty is the most famous example. The predicted bone remodeling due to treatment with a standard stem endoprosthesis is shown in Figure 3.2. At the left of the figure the bone mass density distribution in the immediate postoperative state is depicted for a frontal plane cut and for horizontal slices, where the remaining structure of the natural bone has not changed at all. In this gray-scaled presentation, the increasing bone mass density is indicated from bright to dark color, with the exception of black, which indicates regions of vanishing bone mass. The computed biomechanical equilibrium state is depicted at the right of the figure, and indicates a clear but drastic change in bone mass density in the cortical regions surrounding the stem. This result is in agreement with the socalled ‘‘stress-shielding theorem’’, which postulates that bone resorption occurs in unsuitably stressed domains. These results are underlined by clinical observations, with the predicted bone loss being judged as a referable complication in revisionary treatment. Despite these promising results, the computational techniques available today enable only qualitative prognoses to be made. It is, however, possible to rate


Fig. 3.3 Upper: Simulated time series of bone remodeling caused by a novel metaphyseal-anchored hip-joint endoprosthesis. Lower: Radiographs of a follow-up series.

more or less biomechanically compatible implant designs. In Figure 3.3, for example, the computationally predicted bone remodeling behavior of a new prosthesis model, which has been designed for the treatment of younger patients with a high risk of revision, is compared with X-radiographs from a follow-up series. The computational results clearly show, in qualitative terms, the same tendencies, where the focus is laid on the clearly signed mass density evolution at the tip of the prosthesis. Overall, a stable osseointegration is concluded from the computational results. Additionally, by comparison with Figure 3.1, the innovation is clearly signed, and the bone stock in distal regions remains unchanged because these regions are unaffected by this metaphyseal-anchored device. Therefore, this prosthesis provides a good basis for a revision using traditional techniques. As outlined in this section, computational methods based on phenomenological (but physically consistent constitutive) assumptions are available for the qualitative prediction of the remodeling behavior of bones as result of altered mechanical conditions. Today, these techniques can be used to help identify optimal implant designs in terms of their biomechanical compatibility. In the near future, a supportive system for the treatment of individual patients might become available, providing surgeons with clear advice as to the best choice of available implant, the surgical strategy, and the rehabilitation treatment. At present, although such quantitative predictions are not yet possible, it is not possible to validate these models by using in-vivo measurements. One open question that arises with this macroscopic length-scale relates to the mechanism of stimulation. In the case of this phenomenon – which we refer to

41

42


as mechanosensation and mechanotransduction [4] – the nature of the responsible bone cells must be considered. The problem may be solved in part by using a computational multi-scale analysis, a first approach for which is outlined in the following section.

3.3 A Computational Multi-Scale Approach for Cortical Bone

The overall functional behavior of bone must be explained on several lengthscales because of its hierarchical architecture. Distinctions between spongious and cortical bone are observable at the macroscopic length-scale, with the former being found in metaphyseal regions and short bones such as vertebra, where it consists of a porous, framework-like structure. In spongious bone, the orientation of the rods and their density is clearly correlated to the mean stress trajectories. From the mechanical point of view, the duty of the spongious bone is to provide smooth load transition between joints. Some micromechanical approaches to compute the average mechanical properties and the stress-driven adaptation of spongious bone have been reported (e.g., [11, 25, 26]), and therefore investigations into this topic appear to be well advanced. Cortical bone has a dense structure and is used to build the tube-like bones. When observed microscopically, it is constructed from an amorphous matrix which is reinforced by cylindrical structures, known as osteons or Haversian systems. The osteons are composed of concentric layers of the basic bone material, which is a composite of collagen matrix and hydroxyapatite crystals. The basic composite is organized in fibrils and fibers, which in each layer of an osteon build up a more or less homogeneous but orthotropic structure. The orientation of fibrils changes from layer to layer; thus osteons must be regarded as cross-ply constructions. This hierarchical organization is an essential ingredient for the overall mechanical resistance of bones. This specific construction of osteons is very important from a biological point of view. Between the concentric layers bone cells are embedded the osteocytes; these cells are thought to be responsible for mechanosensation, although the mechanisms involved in this functional flow are the subject of controversy (see [2, 3, 7, 27, 28]). However, knowledge concerning mechanical load transfer in those cross-ply systems can help to explain the local shear stress amplification that is needed by the cells for load detection. Computational methods based on well-established theorems from natural science can contribute to a deeper understanding of this biological self-organizing process. Knowledge of these processes – which without doubt are driven by mechanical stimulation – will be valuable for medical treatment, and they may also open the door to goal-oriented drug treatment. A first approach, highlighting the capabilities of the current state of computational techniques in this field, is presented in the following section.


Fig. 3.4 Schematic diagram of the overall closed-loop control circuit on the osteonal dynamics in cortical bone.

3.3.1 Closed Nano-to-Meso Control Circuit Approach

In order to provide and develop a better understanding of the mechanically driven growth and adaptation of bones, a closed-loop multi-scale algorithm (see Fig. 3.4) has proved invaluable [29]. At the largest length-scale, a section of cortical bone is modeled to include roughly discretized models of osteons (for a detailed description, see Section 3.3.4). The cortical section is loaded by axial compression. At the next smaller length-scale (micrometer scale), detailed finite element models of each individual osteon are created (see Section 3.3.3) which are loaded by the displacement conditions obtained from the cortical section analysis. Each layer of the osteon is modeled with orthotropic material properties that are computed at a subcellular length-scale which depends on the grade of mineralization (see Section 3.3.2). The osteon models are processed for two purposes: (i) to compute the homogenized material properties used in the analysis of the cortical section; and (ii) to compute the mechanical stimulus based on the strains detected by the cell models embedded between the lamellae. Depending on the local strain detected by the cell models, a decision tree is activated for ongoing mineralization, the growth of existing osteons in height, or the creation of new osteons. A sequence of results from this simulation procedure is shown in Figure 3.5. The simulation started with one single osteon in a partially mineralized state. The stage of mineralization is depicted with gray-scales, where black indicates complete mineralization with a 70% hydroxyapatite content. The use of this se-

43

44


Fig. 3.5 Mechanically driven development and mineralization of osteons in a cortical bone section. The grade of mineralization is visualized by the gray scale.

quence illustrates how new osteons are created, how they grow in height, and how the mineralization process occurs continuously. It must be emphasized that from a computational point of view, this multiscale algorithm is easily scalable, because each osteon can be treated independently; for example, each one can be placed on a single personal computer in a network. However, the solution of the overall problem becomes more expensive with increasing density of osteons within a cortical section. The treatment of a complete bone within this multi-scale approach remains visionary, as scale bridging between these models is not yet available. However, the micro-mechanical investigations discussed here may provide a deeper insight into the biomechanical interactions and help in the identification of more reliable constitutive properties for macroscopic approaches. 3.3.2 Sub-Cellular Length-Scale

The basic constituents of bone are collagen molecules reinforced by hydroxyapatite crystals, which form fibrils and fibers. A number of analytical approaches derived from the theory of elasticity can be applied in order to monitor the effective linear-elastic behavior of such a composite. An example is the Mori–Tanaka technique [30], which represents a lower bound for the effective elastic coefficients. When using this method approach it is assumed that the shape of the hydroxyapatite crystals can be approximated by ellipsoids which are embedded in a homogeneous collagen matrix. The obtained elastic properties of the mixture are depicted in Figure 3.6 for the orthotropic mean axis system, depending upon the grade of mineralization. The values for fully mineralized bone agree well with those reported elsewhere [8]. At this point it should be emphasized that the volume fraction of the hydroxyapatite


Fig. 3.6 Young’s modulus of bone tissue in the orthotropic mean axis depending on the grade of mineralization, as computed using the Mori–Tanaka approach. The mean axes have been chosen according to the typical size of hydroxyapatite crystals as a1 ¼ 10 nm, a2 ¼ 2 nm, and a3 ¼ 1 nm.

phase is variable when using this approach, and therefore the time-dependent grade of mineralization of bone-tissue can be represented. 3.3.3 Micro-Scale Model (Single Osteon)

Single osteons, when analyzed at the micrometer-length-scale, show the finite element models to consist of several concentric orthotropic layers, with the mean direction of the transversal isotropy changing from layer to layer. Osteocytes are modeled by assigning randomly distributed elements with isotropic, but much softer, material properties. Due to the orthotropic properties of the layers (which are obtained by using the approach described above) and the embedded osteocyte elements, axis-symmetry does not exist. However, the computational effort can be reduced by the analysis of a representative section with periodic boundary conditions. These computations clearly indicate the strain-amplification caused by the kinematics of orthotropic cross-ply constructions. The absolute values that have been computed even when using this geometrically simple approach correlate well with those obtained from cell experiments [3]. 3.3.4 Meso-Scale Model of Cortical Bone

On the next (centimeter) length-scale, a small section of cortical bone saturated with osteons is modeled by using a finite element approach. At this length-scale

45

46


the osteons cannot be modeled as detailed (see above), and therefore a further step of homogenization must be carried out. Computational techniques are preferred in order to identify the homogenized elastic properties of a cylindrical cross-ply (see [31]). A representative volume element (RVE), which is cut from the osteon model described in Section 3.3.3, serves for the computation of elastic properties observed on the millimeter-length-scale. Besides homogenization, an additional challenge arises with regards to geometric modeling and meshing techniques. The general goal is to compute the dynamic behavior of cortical bone, namely of the creation and growth of new osteons. Thus, at each time step a new geometry and mesh must be created.

3.4 Conclusions

In this chapter, we have reviewed the numeric simulation of stress adaptive boneremodeling phenomena, and outlined the computational methods aimed at optimizing medical treatment and further investigations. A continuum model based on phenomenological assumptions regarding boneremodeling processes was presented, with special emphasis placed on the consistent formulation within the framework of the constitutive theory of continuous media and biomechanical equilibrium. With regards to mechanical loading, the conditions due to joint-forces and muscle interaction and statically equivalent load sets have been computed, based on CT-data and using an inverse simulation technique. The ability of these methods to predict, in qualitative terms, bone remodeling has been demonstrated with a full 3-D analysis of hip-joint endoprostheses. Today, such computations enable distinctions to be made between (biomechanically) compatible implants, and for the identification of optimal prosthesis designs. These computations also help surgeons to choose the best prosthesis, to study the effects of the surgery itself, and to optimize remobilization treatment. Although the continuum modeling approach mainly reflects clinical observations, several uncertainties persist. One problem relates to the constitutive parameters of bone tissue, and another to the mechanism of mechanosensation and mechanotransduction, though neither problem can be resolved at the macroscopic length-scale due to the hierarchical architecture of bone. Today, much effort is expended into modeling approaches on smaller lengthscales, including computations of spongious bone portions [11, 25] and stress analyses on single-cell models [32] to a point where the mechanical responses of single proteins are analyzed [6]. Whilst these investigations are important, a single-scale analysis will not suffice to explain the complex biomechanical interactions involved; this also touched on experimental investigations on cells, and from which the importance of fluid shear on cell stimulation was deduced [2, 5]. Computational methods can help to bridge these gaps, as proposed by the firstorder approach (see Section 3.2), whereby a closed-loop controlled multi-scale computational technique was introduced for the stress-driven osteonal develop-

References

ment of cortical bone, with three typical length-scales being encountered. On the smallest scale, a constitutive model of the composite, comprising collagen molecules and hydroxyapatite crystals, was introduced and served for the analysis of mechanical responses of single osteons, with their time-dependent grades of mineralization. A further homogenization step was needed when analyzing cortical bone in which, driven by the mechanical environment, new osteons are created, grow in height, and become mineralized with time. This overall process is controlled by the strains that are detected at elements and which simulate the mechanical behavior of osteocytes embedded between the orthotropic layers of the osteons. Although this first-order approach to simulating bone growth and aging by bridging several length-scales does not reflect the complicated biology in detail, it does demonstrate the ability of computational mechanics in developing a deeper understanding of these processes. It is clear that, as our computational resources increase, much more detail will become available for each particular length-scale, with the provision of more realistic simulations.

References 1 J. Wolff, Das Gesetz der Transformation

2 3 4

5 6

7 8

9

10 11

12

der Knochen. Hirschwald, Berlin, 1892. E.H. Burger, J. Klein-Nulend, FASEB J. 1999, 13, 101–112. H. Frost, Bone Min. 1992, 19, 257– 271. S.C. Cowin (Ed.), Bone Mechanics Handbook. CRC Press, Boca Raton, 2001. T.D. Brown, J. Biomech. 2000, 33, 3–14. H. Huang, R.D. Kamm, R.T. Lee, Am. J. Physiol. Cell Physiol. 2004, 287, C1–C11. V. Vogel, Annu. Rev. Biophys. Biomol. Struct. 2006, 35, 459–488. J.Y. Rho, L. Kuhn-Spearing, P. Zioupos, Med. Eng. Phys. 1998, 20, 92–102. G.S. Beaupre, T.E. Orr, D.R. Carter, J. Orthop. Res. 1990, 8, 662– 670. M. Doblare, J.M. Garcia, J. Biomech. 2002, 35, 1–17. R. Ruimerman, P. Hilbers, B. van Rietbergen, R. Huiskes, J. Biomech. 2005, 38, 931–941. C.R. Jacobs, M.E. Levenston, G.S. Beaupre, J.C. Simo, D.R. Carter, J. Biomech. 1995, 28, 449–459.

13 C.R. Jacobs, J.C. Simo, G.S. Beaupre,

14

15 16

17 18

19

20

21 22 23

D.R. Carter, J. Biomech. 1997, 30, 603–613. N. Krstin, U. Nackenhorst, R. Lammering, Techn. Mech. 2000, 20, 31–40. D.R. Carter, W.C. Hayes, J. Bone Joint Surg. 1977, 59, 954–962. H. Weinans, R. Huiskes, H.J. Grootenboer, J. Biomech. 1992, 25, 1425–1441. J.C. Rice, S.C. Cowin, J.A. Bowman, J. Biomech. 1988, 21, 155–168. U. Nackenhorst, Proceedings, International Conference on Computer Methods in Mechanics (CMM05), June 21–24, 2005, Czestochowa, Poland. G. Bergmann, G. Deuretzbacher, M. Heller, F. Graichen, A. Rohlmann, J. Strauß, G.N. Duda, J. Biomech. 2001, 34, 859–871. G. Duda, M. Heller, G. Bergmann, Theoret. Issues Ergonom. Sci. 2005, 6, 287–292. K.J. Fischer, C.R. Jacobs, D.R. Carter, J. Biomech. 1995, 28, 1127–1135. B. Ebbecke, PhD Thesis. IBNM, University of Hanover, 2006. U. Nackenhorst, Techn. Mech. 1997, 17, 31–40.

47

48

3 Biomechanics of Bones: Modeling and Computation of Bone Remodeling 24 J.A. Simoes, M.A. Vaz, S. Blatcher, M.

25

26

27 28

Taylor, Med. Eng. Phys. 2000, 22, 453– 459. T. Adachi, K. Tsubota, Y. Tomita, S.J. Hollister, J. Biomech. Eng. 2001, 123, 403–409. T.M. Keaveny, E.F. Morgan, G.L. Niebur, O.C. Yeh, Annu. Rev. Biomed. Eng. 2001, 3, 307–333. L.A. Taber, Appl. Mech. Rev. 1995, 48, 487–545. T.H. Smit, E.H. Burger, J. Bone Miner. Res. 2000, 15, 301–307.

29 C. Lenz, PhD Thesis. IBNM,

University of Hanover, 2005. 30 T. Mori, K. Tanaka, Acta Metall. 1973,

21, 571–575. 31 T.I. Zohdi, P. Wriggers, Introduction

to Computational Micromechanics. Spinger, 2004. 32 J.G. McGarry, J. Klein-Nulend, M.G. Mullender, P.J. Prendergast, FASEB J. 2004, 10.1096/fj.04-2210fje (express article).

49

4 Direct X-Ray Scattering Measurement of Internal Stresses and Strains in Loaded Bones Stuart R. Stock and Jonathan D. Almer

Abstract

High-energy X-ray scattering offers a unique, non-destructive method for quantifying stress in the interior of bones during in-situ loading. The mineral phase and collagen phase of the composite material bone can be studied independently using wide angle X-ray scattering (WAXS or diffraction) and small angle X-ray scattering (SAXS), respectively. In this chapter, X-ray scattering procedures and stress determinations are briefly reviewed, after which the methods used for the studies are summarized and data from several loading experiments presented. Herein, two main results are featured: (i) an independent determination of Young’s modulus in the mineral phase and in the collagen phase of bone via in-situ loading, and comparison with the composite modulus derived from an attached strain gage; and (ii) stress gradients studied in an inhomogeneously loaded rat tibia. Key words:

stress, strain, X-ray scattering, synchrotron radiation, bone.

4.1 Introduction

Most measurements of stress in bone have been made with attached strain gages, and include some human and animal in-vivo studies (for a summary, see [1]). However, in practical terms the number and spacing of such attached gages are very restricted and the data limited to the bone surface. While many remain under the mistaken impression that X-rays are only useful for studying thin surface layers in bone, or for bones reduced to powder, scattering methods employing high-energy X-rays are, in fact, a very attractive means of quantifying internal strains and the related stresses in the interior of mineralized tissues such as bone. High-energy synchrotron X-radiation sources such as the Advanced Photon Source (APS) provide photons that can penetrate millimeters of mineralized Handbook of Biomineralization. Edited by M. Epple and E. Baüerlein Copyright 8 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ISBN: 978-3-527-31806-3

50

4 Direct X-Ray Scattering Measurement of Internal Stresses and Strains in Loaded Bones

tissue. Most neutron and X-ray-scattering studies of bone, a composite of collagen reinforced with a high density of carbonated apatite (cAp) nanocrystallites, have concentrated, however, on the crystalline texture or aspects of crystal quality or stoichiometry (see [2] and references therein), and not on internal stress and strain measurements. In this chapter we describe how high-energy X-ray scattering (E > 60 keV) can be used in situ to measure internal stress in intact bones under applied stress. In this method, wide-angle X-ray scattering (WAXS) or diffraction is used to measure the cAp response to applied stress, and small-angle X-ray scattering (SAXS) to monitor collagen response. The results obtained at the APS illustrate methods of data collection and analysis, and suggest profitable directions for future research.

4.2 Background 4.2.1 X-Ray Scattering

X-rays scatter from the electron clouds of atoms and from nanoparticles and fibrils. An assembly of scatterers with characteristic size or spacing d reinforces scattered intensity in specific directions according to Bragg’s well-known relationship, l ¼ 2dhkl sin y, where l is the X-ray wavelength, dhkl is the crystal lattice spacing, and 2y is the angle between the incident and diffracted beam directions. For bone, the collagen D-period (@67 nm) along the fibril axis produces SAXS peaks for scattering vectors q ¼ 2p=D, and the Angstrom-level periodicities of apatite crystallites produce diffraction peaks in the WAXS regime. Specimens containing many small crystallites with different crystal axis orientations produce cones of diffracted intensity for monochromatic X-rays; Debye cones from different hkl exist simultaneously and produce rings of increased intensity on area detectors [3]. Force applied to a specimen distorts the unit cells and alters the Debye cones (Fig. 4.1). Hydrostatic stresses (those with equal magnitude in all directions) uniformly alter the diameter of cones, whereas deviatoric stresses (those with directionality) change the shape of diffraction rings. Most diffractometers for polycrystalline specimens are equipped with Cu X-ray tubes and utilize 8 keV photons (Cu Ka line). Only @1% of 8 keV X-rays are

Fig. 4.1 Schematic showing: (A) changes in unit cell dimensions under compression; and (B) the corresponding distortion of Debye rings. The solid lines show the unit cell (Debye rings) prior to loading; the dashed lines show the situation after compression.

4.3 Methods

transmitted through 400 mm of cortical bone [4] and, as most researchers are only familiar with standard diffractometers, it is not surprising that many do not realize that options exist for collecting diffraction patterns from the interior of intact, centimeter-sized bones. High-energy photons, available at synchrotrons such as the APS, provide good transmission through bone (at 60 keV there is 10% transmission through @14 mm of cortical bone [4]). 4.2.2 Strains and Stresses

X-ray scattering measures quantities such as dhkl in cAp or the D in collagen, and changes in these quantities define the internal strain imposed during loading – that is, the strain in cAp is ecAp ¼ ðd d initial Þ=d initial and in collagen is ecollagen ¼ ðD Dinitial Þ=Dinitial . Internal stress is a quantity derived from internal strain, and stress sij and strain e kl are second-rank tensors related through the fourth-rank elastic constants Cijkl (i.e., sij ¼ Cijkl e kl ). For a single crystal, the numerical calculation of stress components from strains is straightforward. For data obtained from crystallites with different orientations (i.e., polycrystalline samples), the strains measured by X-rays are averages, and to determine an average stress from these strains requires average elastic constants derived from Cijkl according to one of several approximations. The Reuss approximation assumes that all crystallites experience the same stress; the Voigt approximation assumes that all grains within the sample are subjected to the same uniform strain. The Kro¨ner–Eshelby limit, which is calculated for anisotropic precipitates coupled to an isotropic matrix, yields values of elastic moduli close to those observed experimentally; that is, near the mean of the Reuss and Voigt limits [5]. The Kro¨ner constants are used below.

4.3 Methods 4.3.1 Specimens and Geometry

Figure 4.2 illustrates the transmission diffraction geometry for WAXS and SAXS of loaded bone specimens as is typically used at station 1-ID, APS. The use of area detectors and high-energy X-rays (80.7 keV) and associated small y allows the scattering from the loading (x2 ) and transverse (x1 ) directions to be collected simultaneously. Two X-ray area detectors were used for rapid data collection: a CCD detector for SAXS positioned over 4 m from the specimen and an image plate (IP) detector for WAXS just over 1 m from the specimen. These separations produced adequate SAXS resolution and WAXS angular range, including the cAp 00.2, 22.2 and 00.4 reflections. X-ray exposure times were typically 5 s and 1 s for

51

52


Fig. 4.2 (A) Schematic of the experimental set-up for SAXS and WAXS data collection. Directions x2 and x1 (along the loading direction and transverse to it, respectively) are indicated. (B) MicroCT-derived 3-D image of a 10-mm section of canine fibula loaded in situ. A typical WAXS pattern is shown in the same orientation as the specimen; the lighter pixels on the pattern indicate higher diffracted intensities.

WAXS and SAXS, respectively. Capturing full Debye rings in WAXS required that this detector be centered on the incident beam, and SAXS patterns could not be recorded unless the WAXS detector was translated out of the beam along xIP , an operation requiring @30 s to complete. Both detectors were nominally normal to the incident X-ray beam. Loading in three- or four-point bending is frequently used to study the mechanical properties of rodent long bones [6], but this is less than ideal for high-energy diffraction studies. The large, rapidly varying stress gradients, finite X-ray beam dimensions and averaging over the X-rays’ path through the specimen all degrade the sensitivity of WAXS and SAXS analyses when bending geometries are employed. Therefore, the authors’ experiments have employed uniaxial compression of bone. In order to attach entire long bones or bone sections to the load frame, the ends of the specimens are cast in stiff, water-resistant plastic cylinders (Fig. 4.2) that fit snugly into recesses in the ends of the load train. Small strain gages (one or two per specimen) are glued to the specimen surface so that the macroscopic strain in the gage volume emacro can be compared with those in the two phases of the composite (i.e., ecAp and ecollagen ). The screw-driven load frame used here was constructed at APS for high-energy X-ray scattering determinations of internal stresses, and the details are described elsewhere [2]. The frame’s load cell measures the force applied to the specimen, and laboratory microcomputed tomography (microCT) is used to measure the bone cross-sectional area for use in computing the applied stress, sapplied . Careful calibration (specimen–detector separations, detector tilts relative to the incident X-ray beam) is essential for accurate internal stress/strain analysis. Reference samples with well-defined scattering peaks are used for this purpose (ceria NIST Standard Reference Material SRM-674a for WAXS, and silver behenate for SAXS). Accuracy is further improved by using a laser distance gage to correct for small specimen shifts during loading and for curvature of the bone surface.

4.3 Methods

Fig. 4.3 (A) WAXS pattern showing definition of aximuthal angle h and radius from the pattern center. The darker the pixel, the higher the diffracted intensity. (B) Experimental 00.4 peak position versus azimuthal angle for a section of canine fibula at the indicated applied compressive stresses (MPa). The quantities r and h are defined in the text. (Figures from Ref. [2].)

4.3.2 Analysis of Two-Dimensional (2-D) Scattering Patterns

The scattering pattern analyses are performed using FIT2D [7] and MATLAB programs developed at Sector 1 of APS. Figure 4.3 shows a typical 2-D WAXS pattern and the variation of 00.4 peak position versus azimuthal angle h for several applied compressive stresses, but this is not visible at the scale shown. In order to see any differences, each Debye ring is divided into 72 azimuthal bins, and intensity versus radius is plotted. The resulting 1-D radial plots are fitted with pseudoVoigt functions to provide radial peak position rh that were converted into absolute lattice plane spacings d h using known ceria d-spacings. For the experimental conditions of Figure 4.2, estimates of absolute error in d h are below 104 [8]. For each reflection studied, the profiles of radius versus h obtained at the different applied stresses intersected at a single radius, the invariant radius r , and the corresponding azimuthal orientation was h (see Fig. 4.3B). Measured radii were referred to this r value to provide orientation-dependent (deviatoric) strain values: eh ¼ ðrh r Þ=r . These strain profiles, eh , were fitted to a biaxial strain model [9] to account for sample geometry, and provided values of the deviatoric strain components exx and eyy . These strain components, along with the X-ray elastic constants, were then used to calculate stress along the loading direction (see Section 4.3.4). The WAXS patterns show azimuthal intensity variations – that is, diffracted intensity around the 00.2 Debye ring varies substantially. In the case of the long bones described here, the diffraction patterns show that the cAp crystallites have c-axes oriented primarily along the axes of the long bones, which was an expected

53

54


Fig. 4.4 (A) SAXS pattern from mineralized turkey tendon. The lighter the pixel, the greater the scattered intensity. The black arrows point to 3rd and 5th order D-period peaks, and the longitudinal (L) and transverse (T) directions for the tissue are shown. (B) SAXS intensity as a function of scattering vector q ¼ 2p/D for the pattern shown in using G9 azimuthal integration to improve the signal-to-noise ratio. Data along the tendon L and T directions are shown.

result. Although cAp texture in bones is not the subject of this chapter, it is important to realize that extreme texture produces incomplete Debye rings and can seriously interfere with curve fitting (e.g., of the data in Fig. 4.3B). Crystallite size and microstrain analyses can also be performed using the radial shapes of the diffraction peaks in WAXS patterns, such as Fig. 4.3A (see [2]). Analyses can be as simple as the peak width analysis of the Williamson and Hall method (see [3]), or as complex as the peak shape analysis of the Warren– Averbach method (see [5]). Figure 4.4 shows typical SAXS patterns from cAp-mineralized tissue; this example is from hydrated turkey tendon. A 2-D pattern is shown in Figure 4.4A, while Figure 4.4B shows I(q) plots for transverse T and longitudinal L tendon directions. At least 12 peaks are seen, and the use of multiple peak positions (on both sides of the incident beam) improves the precision with which D can be determined. In Figure 4.4, D ¼ 67.4 nm, which is about the value expected. It should also be noted that if no mineral were present, the peak intensities would be much weaker. Variation of peak intensity with order (here, the odd-order peaks are more intense than the even) can be used to determine the fraction of D-period occupied by mineral [10]. This type of calculation is fairly straightforward, and is used not only in SAXS analysis but also in the analysis of structures such as multiple quantum well structures grown by molecular beam epitaxy [11]. The shape of the I(q) curve along the transverse direction can be analyzed to provide information concerning mineral crystallite size and shape [12].

4.4 Examples of Data and Analysis

4.3.3 X-Ray Elastic Constants and Strain–Stress Conversion

For the biaxial geometry used in experiments described below, the transverse strains exx and ezz are assumed to be equal, and the stress and strain tensors are related through [13]: syy ¼

1 S1 eyy ðeyy þ 2exx Þ ; S2 =2 S2 =2 3S1

ð1Þ

where S1 ðhk:lÞ ¼ ðn=EÞaverage and S2 =2ðhk:lÞ ¼ ðð1 þ nÞ=EÞaverage are the average X-ray elastic constants [5] computed from values of Cijkl using the Kro¨ner– Eshelby model via the computer program Hauk.exe (available from the authors upon request): this mathematical formulation is too lengthy to reproduce here [13, 14]. In other words, measured strains exx and eyy (where x and y are along the horizontal and vertical axes of the 2-D diffraction pattern, respectively, and the X-ray beam is along z) are used to calculate internal stresses along the loading (i.e., y) direction. The change of notation (exx for e11, etc.) emphasizes the coordinate axes of the detector as opposed to the specimen.

4.4 Examples of Data and Analysis

Measurement of the strains in cAp and collagen (from WAXS and SAXS, respectively) as a function of sapplied (measured by the load cell of the mechanical testing apparatus) allows one to determine Young’s modulus for the individual constituent phase of bone. The strain gage data provides a measure of Young’s modulus (i.e., the macroscopic value). Figure 4.5A shows ecAp , ecollagen and emacro as a function of sapplied for a section of canine fibula kept in a hydrated state by dripping phosphate-buffered saline onto the specimen’s surface. The points for the cAp phase are the average of values for the 22.2 and 00.4 reflections. The resulting moduli (90% confidence limit) are: Emacro ¼ 24:7ð0:2Þ GPa, EcAp ¼ 41ð1:0Þ GPa, and Ecollagen ¼ 18ð1:2Þ GPa [15]. The value for Emacro is in good agreement with moduli of similar bone types reported elsewhere. The modulus for cAp is about one-third of that of inorganic apatite [16], and this presumably reflects the nanocrystalline form of cAp found in bone in intimate contact with collagen. The value of Ecollagen is at least nine times higher than one would expect [17, 18]; even though the collagen D-periodicity is the basis for this determination, most of the scattering power is from the templated cAp, and straight-jacketing of the collagen [18] may be the reason for the large experimental modulus. The collagen straight-jacket model treats bone as a composite of a very high-volume fraction of stiff rods (mineralized collagen fibrils which themselves are nanoparticulate reinforced structures) glued together by a low-volume fraction of non-collagenous

55

56


Fig. 4.5 (A) Longitudinal strain measured in cAp (WAXS), in collagen (SAXS) and by the strain gage (macroscopic) as a function of applied compressive stress for a 10-mm-long section of canine fibula. The WAXS data are the average for 22.2 and 00.4 reflections. (Plot is adapted from Ref. [15].) (B) Longitudinal strain as a function of applied

compressive stress in a rat tibia. A 3-D rendering of the specimen is inset at the lower left, and data from the strain gage are shown as well as that for the 00.2 cAp reflection measured at five positions across the specimen (along x). The strain gage data is offset by 7 104 e (i.e., vertically) for clarity.

proteins; this is quite different from a simple, uniform, discontinuously reinforced composite. From this perspective, what is measured by SAXS probably reflects the stiff rod response, whilst what is measured by WAXS reflects the cAp response and Emacro is the combined response of rods and inter-fibrillar mineral and proteins. Figure 4.5B shows data illustrating how the high-energy diffraction methods can be used to investigate stress gradients in bones, or assemblies of bones. Strain versus sapplied is shown for a rat tibia loaded in compression. The intrinsic curvature of this long bone results in a significant amount of bending when the bone is compressed, a situation that mimics the situation in vivo. The bone’s axis of curvature was positioned parallel to the incident X-ray beam direction so that simple lateral translation of the specimen allowed volumes under different stress states to be probed. Data from the strain gage and from five positions across the specimen are shown in Figure 4.5B.

4.5 Discussion and Future Directions

It should be noted that others have used diffraction to measure stresses in thin sections of bone [19], and recent studies combining SAXS with in-situ loading of thin bone specimens have contributed much to our understanding of bone deformation [20, 21]. The studies summarized in this chapter focus on intact bone cross-sections (canine fibula) or entire bones (rat and mouse tibiae), which is con-

References

siderably different from thin-section studies. The results presented herein show that it should be quite simple to extend the approach to undissected assemblies of bones and to the in-vivo loading of animal models, such as the rodent ulnar [22] or tibial [23] loading models. It is difficult – if not impossible – to obtain these data by other means.

Acknowledgments

The authors thank Dr. W. Landis and his group (Northeastern Ohio University College of Medicine) for providing the mineralized turkey tendon, Dr. R. Sumner and his group (Rush Medical College) for providing the canine fibula, and Dr. K. Igarashi (Tohoku University, Japan) for providing the rat tibiae. Use of the APS was supported by the US Department of Energy, Office of Science, Office of Basic Energy Science, under contract No. W-109E-ENG-38.

References 1 S.P. Fritton, C.T. Rubin, in: S.C.

2 3

4

5

6

7

Cowin (Ed.), Bone Mechanics Handbook, 2nd edn. CRC Press, Boca Raton, 2001, pp. 8.1–8.41. J.D. Almer, S.R. Stock, J. Struct. Biol. 2005, 152, 14–27. Fundamentals of X-ray scattering and diffraction appear in texts such as: B.D. Cullity, S.R. Stock, Elements of Xray Diffraction, 3rd edn. Prentice-Hall, New York, 2001. NIST, July 2001. Tables of X-Ray Mass Attenuation Coefficients and Mass Energy Absorption Coefficients: from 1 keV to 20 MeV for Elements Z ¼ 1 to 92 and 48 Additional Substances of Dosimetric Interest, NISTIR 5632. I.C. Noyan, J.B. Cohen, Residual stress: Measurement by diffraction and interpretation. Springer, New York, 1987. C.H. Turner, D.B. Burr, in: S.C. Cowin (Ed.), Bone Mechanics Handbook, 2nd edn. CRC Press, Boca Raton, 2001, pp. 7.1–7.35. (a) A.P. Hammersley, S.O. Svensson, A. Thompson, Nucl. Instrum. Methods 1994, A346, 312–321; (b) A.P. Hammersley, S.O. Svensson, M. Hanfland, A.N. Fitch, D. Haüser-

8

9

10

11

12

13

14 15 16

mann, High-Press. Res. 1996, 14, 235– 248. J. Almer, U. Lienert, R.L. Peng, C. Schlauer, M. Odeń, J. Appl. Phys. 2003, 94, 697–702. B.B. He, K.L. Smith, in: Society for Experimental Mechanics Annual Conference and Exposition, Houston, TX, 1998, pp. 217–220. A. Ascenzi, A. Bigi, M.H.J. Koch, A. Ripamonti, N. Roveri, Calcif. Tissue Int. 1985, 37, 659–664. P.C. Huang, S.R. Stock, A. Torabi, C.J. Summers, Adv. X-ray Analysis 1990, 33, 67–74. W. Tesch, T. Vandenbos, P. Roschgr, N. Fratzl-Zelman, K. Klaushofer, W. Beertsen, P. Fratzl, J. Bone Miner. Res. 2003, 18, 117–125. V. Hauk, Structural and residual stress analysis by nondestructive methods: Evaluation, application, assessment. Elsevier, New York, 1997. Computer program Hauk.exe; available from J. Almer, upon request. J.D. Almer, S.R. Stock, J. Struct. Biol. 2007, 157, 365–370. (a) R.S. Gilmore, J.L. Katz, J. Mater. Sci. 1982, 17, 1131–1141; (b) T.N. Gardner, J.C. Elliott, Z. Sklar, G.A.D.

57

58


17 18 19 20

Briggs, J. Biomech. 1992, 251, 265– 1277; (c) M.C. Sha, Z. Li, R.C. Bradt, J. Appl. Phys. 1994, 75, 7784–7787. N. Sasaki, S. Odajima, J. Biomech. 1996, 29, 655–658. I. Ja¨ger, P. Fratzl, Biophys. J. 2000, 79, 1737–1746. K.S. Borsato, N. Sasaki, J. Biomech. 1997, 30, 955–957. H.S. Gupta, W. Wagemaier, G.A. Zickler, D. Raz-Ben Aroush, S.S. Funari, P. Roschger, H.D. Wagner,

P. Fratzl, Nano Lett. 2005, 5, 2108– 2111. 21 H.S. Gupta, W. Wagemaier, G.A. Zickler, J. Hartmann, S.S. Funari, P. Roschger, H.D. Wagner, P. Fratzl, Int. J. Fract. 2006, 139, 425–436. 22 S.P. Kotha, Y.F. Hsieh, R.M. Strigel, R. Mu¨ller, M.J. Silva, J. Biomech. 2004, 37, 541–548. 23 T.S. Gross, S. Srinivasan, C.C. Liu, T.L. Clemens, S.D. Bain, J. Bone Miner. Res. 2002, 17, 493–501.

59

5 Osteoporosis and Osteopetrosis Adele L. Boskey

Abstract

Osteoporosis, a common bone disease which generally affects the elderly, and osteopetrosis, a much rarer disease which appears early in life, share features of defective osteoclast activity, abnormal osteoblast activity, increased tendency to fracture, and altered bone mechanical properties. These features in both cases are associated with geometric and material abnormalities. In this chapter we review these diseases, their molecular and cellular bases, and the mineral and matrix properties as determined by X-ray diffraction and vibrational spectroscopy of bone from humans and animals with these conditions. The features of animal models of each of these conditions are compared with the presentation in humans. Key words: bone formation, bone remodeling, osteoporosis, osteopetrosis, FTIR microspectroscopy, Raman spectroscopy, X-ray diffraction, electron microscopy, osteoclasts, osteoblasts.

5.1 Introduction: Two Distinct Diseases with Common Features

Although patients with osteoporosis or osteopetrosis are all at an increased risk of fracture, these diseases – in clinical terms (see Table 5.1) – are very different. For example, osteoporosis is quite prevalent, and affects four in five women and one in eight 8 men [1]. In contrast, osteopetrosis is much rarer, occurring in from 1 in 200 000 individuals in the United States [2], while estimates from an earlier population-based study in Finland suggested that it occurs in 11 in 200 000 individuals [3, 4]. In terms of patient age, osteopetrosis occurs most often in childhood, although there is a rarer adult form [5]; by contrast, osteoporosis does have some childhood variants [6–8], but it is generally a disease of older individuals [9]. Both, osteopetrosis (which is also referred to as ‘‘marble bones’’ or Handbook of Biomineralization. Edited by M. Epple and E. Baüerlein Copyright 8 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ISBN: 978-3-527-31806-3

60

5 Osteoporosis and Osteopetrosis Table 5.1 Comparison of clinical features of osteopetrosis and osteoporosis.

Clinical feature

Osteopetrosis

Osteoporosis

Skeletal shape

Short, thicker bones

No growth abnormalities, thinner cortices

Bone mineral density

Elevated

Decreased

Tissue connectivity

Increased

Decreased

Fracture incidence

Elevated

Elevated

Failure to heal fractures

High

Low

Albers–Scho¨nberg disease) and osteoporosis have a variety of forms [10]. Moreover, although the processes of mineralization and skeletal maintenance in each condition are distinct, the diseases share common features. In the healthy individual the processes of bone formation (by osteoblasts) and bone remodeling or turnover (by osteoclasts) are coupled, but in both osteopetrosis and osteoporosis this coupling is lost. The distinctive feature of osteoporosis, by definition, is the presence of more-porous (less-dense) mineralized bone with an increased tendency to fracture [1]. Osteopetrosis, in contrast, is characterized by an increased amount of calcified cartilage, which results in extremely dense bones that also tend to fracture [10]. The altered properties of both the mineral and the matrix in each of these conditions contribute to the bone fragility. 5.1.1 Comparisons of Clinical Features of Osteoporosis and Osteopetrosis 5.1.1.1 Histology In the bones of patients with either osteoporosis or osteopetrosis, the osteoclasts – the cells that remove mineral and matrix from bone in response to signals (for reviews, see [11, 12]) – do not function properly. In general, in osteopetrosis the osteoclasts have a decreased activity, whereas in osteoporosis their activities are often increased. In osteopetrosis, the impaired osteoclastic activity results in an accumulation of calcified cartilage (Fig. 5.1a) that, in contrast to the normal situation, is not replaced by bone. Bone formation by osteoblasts is generally also impaired, but in certain cases it may be normal. In osteoporosis, there is generally an imbalance between osteoblastic and osteoclastic activity, with too little new bone formation and too much bone resorption, leading to the loss of connectivity (Fig. 5.1b), which contrasts with the situation in healthy, control bone (Fig. 5.1c).

5.1 Introduction: Two Distinct Diseases with Common Features

Fig. 5.1 Histologic and radiographic characteristics of cancellous bone in (a) osteopetrosis, (b) osteoporosis and (c) normal bone. Note the persistence of calcified cartilage within the woven bone in the osteopetrotic tissue, and the thin struts in the osteoporotic bone. (Reproduced, with permission, from Peter G. Bullough, Orthopaedic Pathology, 4th edn. Mosby, New York, 2004, Figures 7.10a, 7.47, and 7.48a.)

5.1.1.2 Radiography Similar to the two-dimensional (2-D) histology of bone sections, whole-bone 2-D radiographic images of osteoporosis show a loss of bone mass. In contrast, radiographs of the bones of patients with osteopetrosis demonstrate an increase in tissue density due to the accumulated calcified cartilage. Many osteopetrotic patients are dwarfed and anemic, some are blind and deaf as a result of pressure on the relevant nerves, and most have thickened skulls [13]. It is believed that the artist Toulouse-Lautrec was a victim of some form of osteopetrosis [14]. Despite the difference in radiographic appearance when patients with osteopetrosis and osteoporosis are compared, there is an increased risk of fracture in both conditions [4]. In osteopetrosis there is also a high incidence of non-unions (failure to heal fractures), because a key feature of the normal fracture-healing process is the removal of calcified cartilage by osteoclasts, followed by replacement with bone. The fragility fractures in osteoporosis generally occur through the thinning trabeculae, but generally heal without non-union complications.

61

62

5 Osteoporosis and Osteopetrosis

Clinically, bone mineral density (BMD) measurements are used for the diagnosis of both, osteoporosis and osteopetrosis. Two-dimensional values (areal BMD) are calculated from dual photon absorptiometry (DXA); the resultant value is generally reported as a T-score, the standard deviation (SD) of the patient’s BMD compared to a healthy young-adult reference population. The T-score is calculated as: T ¼ (patient BMD mean young adult BMD)/1 SD young adult BMD; with BMD and SD expressed in units of g cm2 .

Z-scores, which compare similarly aged individuals, are less frequently used. In the diagnosis of osteoporosis, individuals with a T-score less than 1.5 are at risk for fracture, individuals with a T-score less than 2.5 are defined as osteoporotic, and those with positive score are considered not at risk of fracture [1]. Patients with osteopetrosis have very dense bones, and high values of BMD, with positive T-scores; such people are also at high risk of fracture. 5.1.2 Comparisons of Bone Mineral Properties in Osteoporosis and Osteopetrosis

Differences have been identified between the mineralization processes and the properties of the bone mineral and matrix in osteoporosis and osteopetrosis when contrasted with age- and gender-matched, disease-free controls. These differences have been elucidated by analyses of the bones themselves, and also from analyses of cell and organ cultures derived from these bones. Reviews of these methods can be found elsewhere [15]. The strength of bones, or their ability to resist fracture, is determined by the amount of mineral present (BMD), the geometry (shape) and architecture of the bones, the composition of both the mineral and the matrix, and the presence of micro-cracks. Methods for evaluating whole-bone properties range from mechanical tests to micro-computed tomography [16], and from histochemistry to in-situ hybridization. X-ray diffraction and chemical analyses, nuclear magnetic resonance (NMR), energy-dispersive X-ray analysis (EDX), and vibrational spectroscopic techniques provide insight into mineral and matrix properties [15, 17]. The vibrational spectroscopy parameters have been validated by comparison with independent methods, and include information on mineral content, mineral crystal size, mineral crystal composition, and matrix maturity [17–19]. The availability of array detectors has enabled the rapid detection of spectra in sections of tissues with a spatial resolution under 7 mm. The calculation of peak area ratios or intensity ratios in these multispectral files permits the generation of hyperspectral images where the x- and y-axes correspond to locations in the tissue and the z-axis to the value of the parameter in question. Using these techniques to assay bone biopsies from patients with these diseases, in osteopetrosis the bones are found to be more dense, and the mineral and matrix to be less mature than in control bones. The crystals in the osteope-

5.2 Animal Models of Osteoporosis and Osteopetrosis

trotic bone tend to be smaller and have more imperfections (inclusions, adsorbed ions, etc.) than those in bones from age- and gender-matched controls. In contrast, in osteoporosis the average mineral and matrix properties in the moreporous bone appears similar to that of bones from older individuals than of the age-matched controls. In other words, osteoporotic bone tends to contain larger, more perfect crystals, with a higher carbonate content and less acid phosphate than control bones. In both osteopetrotic bone and osteoporotic bone, the distribution of mineral properties differs from that seen in the bones of healthy individuals. Most of the detailed information on mineral properties in these diseases has been obtained studies with animal models, and consequently it important to discuss whether these are valid models of the human condition. In the following sections the characteristics of animal models of osteoporosis and osteopetrosis will be compared to the bone properties observed in humans with these conditions. Animal models provide the advantage of being able to compare similar genetic backgrounds, of obtaining larger amounts of tissues for analysis, and allowing controlled studies to be conducted of pharmaceutical therapies.


A variety of naturally occurring drug-induced and surgically induced ‘‘models’’ of osteoporosis [20–65] and osteopetrosis [66–103] are available, in addition to more recently developed genetically engineered mutants (Table 5.2). Although these models all provide insights into the human conditions, they do not all totally resemble the diseases in humans. Thus, in the following discussion it must be recognized that analyses of human tissue, where possible, may yield different information about the mineralization process and mineral properties than would the animal models. This occurs, in part, because the diseases are so heterogeneous, in part because they are influenced by complex factors (multiple genes and genetic interactions, diet, exercise, etc.), and in part because modifying gene expression in the animal may not have the same effect as altering the gene in humans. 5.2.1 Osteoporosis 5.2.1.1 Rodent Models As osteoporosis has long been recognized as a disease which is more prevalent in postmenopausal women and elderly men, the early animal models were based on the ovariectomy and castration of rats [20, 21] and later of mice [22]. Ovariectomized/castrated rats had bones that, during mechanical testing, broke more easily [23], although they did not fracture during normal activities. The immobilization of rodent bones has been used to mimic the bone loss associated

63

64


with bed-rest or space flight in humans. The most popular model – hind-limb suspension – causes a rapid loss of bone in the hind limbs [24]. Similarly, steroid-induced osteoporosis in rodents [25] did not result in spontaneous fractures, yet the bones were weaker, less dense, and had altered mineral properties, especially around the osteocytes [26]. The ovariectomized rat is an accepted model for testing anti-osteoporotic drugs [21], while the ovariectomized mouse provides the additional advantage of being able to modulate gene expression during these studies. However, because in rodents the cortical bones thicken during aging, whilst in humans the cortices become thinner with age, the mineral properties and mineralization processes in rodents may not always be the best model for comparison to human bone. A limited number of rodent models do fracture spontaneously; examples include the accelerated aging mouse (SAM) [27], in which the mechanical properties are distinct from those of age-matched controls [28]; and the spontaneously fracturing mouse, sfx [29], which lacks an enzyme needed to process vitamin C, a cofactor required for collagen synthesis. The SAM mouse can lose up to 60% of its bone mass during aging, [30]. Interestingly, while the long bones of these mice are mechanically weaker, the vertebrae – though reduced in trabecular volume, number, and thickness – do not show the reduced mechanical strength associated with osteopenia in other models, or in humans [28, 31]. The bones of SAM are weak and brittle, the matrix contains less collagen, and the collagen fibrils are not properly organized. However, the mineral properties, as determined by Raman analyses, do not differ from those in controls [32]. Mice in which genes have been ablated (knockouts), inserted (knockins), or otherwise modified (transgenics) while not fracturing, often develop phenotypes with decreased bone density (osteopenia), and these may provide important insights into the mineralization process. For example, the osteonectin knockout loses trabecular bone with age and shows extensive cortical thinning [33]. Osteonectin is a matrix glycoprotein that inter alia regulates collagen fibrillogenesis. An early phenotype of these animals is the development of cataracts [34]. As these animals age, their bones become weaker in torsion (i.e., they fracture more easily) than wild-type controls, and the mineral content of their bones is increased, as is the average crystal size and maturity of the collagen [35]. The increase in crystal size and matrix maturity is similar to what is seen in humans with osteoporosis, but in contrast to the frequently noted decreased mineral content in osteoporotic humans [36]. The biglycan knockout mouse also loses bone with age [37]. Biglycan is a small leucine-rich proteoglycan that regulates collagen fibrillogenesis and binds growth factors within the matrix. In vitro, and in the absence of fibrillar collagen, biglycan can act as a bone mineral (hydroxyapatite) nucleator [38]. The biglycan knockout mouse has weaker bones than its wild-type control, decreased numbers of trabeculae with decreased mineral content, and increased bone mineral crystal size [37]; this situation is more analogous to that seen in osteoporotic humans. The


biglycan knockout phenotype is dependent on the background of the mouse in which the gene is ablated [38]. The above models indicate the importance of collagen organization for the mechanical properties of bone. Humans and animals with osteogenesis imperfecta (OI), or ‘‘brittle bone disease’’, further illustrate this. OI is a rare birth defect which is due, for the most part, to a variety of different mutations affecting the formation of type I collagen, the principal matrix component of bone [39]. The animal models of OI all show increased bone brittleness (reduced energy required to break the bones and decreased elasticity), with the severity and number of spontaneous fractures depending on the nature of the genetic mutation [40– 44]. Some regions of bones of OI patients and animal models have mineral crystals outside the collagen matrix [45]. Additionally, in OI bones there are fewer mineral crystals than in healthy age- and gender-matched bone, and in general the crystals are smaller and have a composition which is distinct from that of the age-matched controls [46, 41]. As in humans with OI, the fracture incidence in mice (if the disease is not perinatal lethal) diminishes with age, which is quite the opposite of what happens in osteoporosis. In fact, in osteoporotic humans the probability of having a second or third fracture after the first fragility fracture is markedly increased during the first year [47]. A number of other transgenic and knockout mice that have osteopenic phenotypes for which mineral properties also exist, but have not been reported. For example, mice that over-express interleukin-4 (IL-4), a cytokine which has multiple effects on a variety of cell types, have decreased osteoblastic activity [48]. These mice (both sexes) develop a ‘‘hump-back’’ with age (reminiscent of the dowager’s hump in osteoporotic women), but the mice bones do not fracture spontaneously. The cortical thickness in the long bones of these mice decreases with age, as did the trabeculae in the vertebral bodies. Histological studies have revealed no evidence of osteomalacia or other bone diseases, but the bones of the transgenics had reduced mechanical strength. Similarly, mice overexpressing noggin (one of many antagonists of bone morphogenetic proteins; BMPs) in bone cells [49] show significantly decreased osteoblastic activity with age. At the age of 8 months these mice had increased marrow space, a significantly lower BMD in all bones, decreased bone formation rates, and decreased osteoblastic and osteoclastic activities [50]. Another BMP antagonist, sclerostin (a product of the SOST gene expressed by osteoblasts and osteocytes), when overexpressed under a bone-specific promoter resulted in mice that had less bone, less mineral within the existing bones, and bones which were mechanically weaker [51]. As reviewed elsewhere, knockout of the high bone density gene, LRP5, or disruption of the factors with which it interacts, also produces osteopenia in mice [52]. A stem-cell antigen cell-surface protein, Sca-1, when knocked out, causes development of osteopenia in older mice associated with a failure to replenish osteoblasts rather than a failure for pre-osteoblasts to undergo osteogenesis [53]. Factors such as osteoprotegrin (OPG), which regulates osteoclastogenesis, when

65

66


knocked out can also provide models of osteoporosis; the OPG-knockout has a decreased bone density [54]. Additional mouse models with osteopenic phenotypes have not been analyzed for mineral properties (for a review, see [55]). 5.2.1.2 Non-Rodent Models Larger animals mimic more closely the loads placed on human bone and, when ovariectomized, they have been used as osteoporotic models. Thus, ovariectomized dogs [56], ewes [57], mini-pigs [58] and non-human primates [59] have been analyzed. Dietary manipulation has also been used, with or without ovariectomy, to mimic human osteoporosis in these animals. In an on-going study, bones from adult sheep treated with a diet to induce metabolic acidosis, ovariectomized sheep, and ovariectomized sheep given the metabolic acidosis diet showed no significant difference in mineral content relative to controls, but all had increased crystallinity, increased carbonate content, and increased collagen maturity relative to controls, similar to the situation seen in humans with osteoporosis. The non-human primates are the only animals that fracture in the wild. Such animals have menstrual cycles, and Haversian systems similar to those in humans. In Caribbean monkeys, cortical mineral density and porosity increased with age, and vertebral density and cortical area increased with animal weight [60]. Ovariectomized cynomolgus monkeys develop accelerated bone loss, increase bone turnover, and have reduced bone strength [59] relative to shamoperated controls. Density fractionation and mineral analyses by X-ray diffraction of the jaws of immobilized monkeys show a significant shift toward higher density fractions indicative of the presence of a greater than normal content of mature highly mineralized bone and a parallel decrease in crystallite size [61]. Young ovariectomized and sham-operated cynomolgus monkeys have less-dense bones than control animals; crystal size in these animals was not altered [62]. By using infrared (IR) microspectroscopy, it was reported [63] that trabecular bone from ovariectomized monkeys had significantly lower mineral-to-matrix ratios, a parameter directly related to ash weight [64] with values of 5.8 G 0.2 compared to controls (6.2 G 0.2; p a 0.05) and contained larger/more perfect apatite crystals (increased crystallinity) with increased carbonate: phosphate ratios. Similarly, using synchrotron-based IR microspectroscopy, Miller et al. [65] found an increased acid phosphate content and different collagen structure in ovariectomized monkeys at 2 years post ovariectomy. Reduced rates of mineralization were also found in these animals. These results are in good agreement with findings in osteoporotic humans, even when taking into account the population variation. 5.2.2 Osteopetrosis

In humans, the osteopetroses are a heterogeneous group of bone-remodeling disorders characterized by an increase in bone density due to defects in bone remodeling (osteoclastic resorption), and with an increased incidence of fracture [66].


These diseases are usually classified based on inheritance, age of onset, severity, and clinical symptoms. They include infantile malignant autosomal recessive osteopetrosis, an intermediate (milder) autosomal recessive form, adult benign autosomal dominant osteopetrosis type I, and autosomal dominant osteopetrosis type II. Another variant, known as pycnodysostosis, has also been reported. This disease is due to a deficiency in cathepsin K activity, and may involve the impaired formation of osteoclasts, the impaired activity of these bone-resorbing cells, or both. Various naturally occurring and genetically modified animal models mimicking different forms of osteopetrosis have been studied in efforts to elicit further understanding of the pathogenesis of the disease and to evaluate potential treatments. In addition, some drugs that block osteoclast action have been found to induce an osteopetrotic phenotype in animals and humans. 5.2.2.1 Rodent Models The first recognized rodent models of osteopetrosis were the mutant animals described by Marks and colleagues [67]. The earliest observed phenotype in many of these rodents is failure of the teeth to erupt, as bone must be removed by osteoclasts for this to occur. While the functional activity of the osteoclasts in these rodents is impaired for a variety of reasons, the resulting bone phenotype is quite similar, with a persistence of calcified cartilage, and the presence of excessive mineralized tissue containing smaller crystals than comparable tissue in normal controls. For example, the toothless rat (tl/tl) [68], has a decreased bone mineral crystal size relative to normal controls. The collagen matrix in these rats is characterized by slight decreases in reducible cross-links and increases in the content of the stable cross-links, pyridinoline, and deoxy-pyridinoline, reflecting the persistence of more mature tissue. The incisor-absent (ia/ia) rat, similarly, has a higher mineral content in both their calvaria and long-bone metaphyses than age-matched controls [69]. The ia/ia mineralized tissue contains smaller mineral crystals relative to controls. The metaphyses in the ia/ia rats also has an elevated hexosamine content, deriving from cartilage proteoglycans, and verifying the persistence of cartilage. A naturally occurring mouse model (op/op) lacks one of the factors essential for osteoclast differentiation [70]. These mice also have decreased bone-forming (osteoblastic) activity. The presence of an altered osteoblast phenotype was also suggested in studies of tl/tl rats [71]. These data suggest that the coupling of osteoblasts and osteoclasts [72] may be disturbed in these models. While these rodent models have increased bone mineral density and cross-sectional geometries relative to their controls, their bones are weaker than their normal littermates, and their cortices significantly thinner [73]. The analysis of a necropsy sample from a human infant with the lethal form of osteopetrosis similarly showed the presence of smaller crystals [74]. In contrast to these models, the gray lethal (gl/gl) osteopetrotic mouse, with the most severe form of osteopetrosis [75, 76], and @10% of patients with malignant infantile osteopetrosis, have a defective bicarbonate-chloride transport system [77,

67

68


78]. This chloride transport is essential for the osteoclast-mediated acidification of the extracellular matrix necessary to remove the mineral, and for the activity of other lysosomal enzymes. In this severe variant, while osteoclast activity is impaired, osteoblast activity is normal, which leads to the suggestion that the chloride transport mechanism might be targeted for the treatment of osteoporosis [79]. Other less-severe rodent models include the op/op mouse that fails to resorb bone due to a defect in macrophage colony-stimulating factor [70], the congenitally osteosclerotic oc/oc mouse that lacks a functional H(þ)-ATPase a3 subunit (also important for acidifying the extracellular matrix) [80], and the microphthalmic mi/mi mouse [81]. The op/op mouse has defective osteoblast activity and impaired mineralization [70], while the small-eyed mi/mi mouse has a defect in the microphthalmia-associated transcription factor (MITf ). These mice have decreased osteoclast formation [81], while their marrow cells over-express receptor activator of nuclear factor kappa B (RANK) ligand (RANKL), a factor which, when binding to its receptor, activates osteoclastogenesis [82]. Another rodent model of osteopetrosis mimics the condition known as pycnodysostosis, which is linked to a defect in cathepsin-K, a lysosomal cysteine protease elevated in activity in active osteoclasts [83]. Detailed structural analyses of the bones of one young patient and one older patient with this disease showed thickening of the bone mineral particles, poorly aligned crystals associated with collagen fibrils, and abnormal trabecular structure [84]. The cathepsin K knockout mouse [85] had abnormal matrix turnover, but no histochemically apparent defects in either mineral removal or accretion. Infrared imaging of the bones of cathepsin K null animals (kindly provided by Drs. Gelb and Schaffler of Mount Sinai Medical School) showed not only the persistence of highly mineralized calcified cartilage, but also a decreased crystallinity and decreased matrix maturity similar to that reported in a limited number of human biopsies (Fig. 5.2). These mice also show disorganized matrices and increased bone fragility associated with increased osteoclast recruitment [86]. 5.2.2.2 Other Osteopetrotic Models While bovine models of osteopetrosis have been reported [87–89], and avians infected with viruses develop an osteopetrosis-like disease [90, 91], only the bovine mechanical properties have been shown to be similar to those of humans [88]. Mineral properties have not been characterized. Other models that have an osteopetrotic phenotype, most often associated with genetic modifications leading to impaired osteoclast activity, are listed in Table 5.2. Mineral properties in the osteocalcin knockout [92] demonstrated the presence of smaller crystals, increased calcified cartilage, and increased bone mass, like that found in osteopetrotic models. The only other knockout in which mineral properties have been described is the c-fos knockout [95], which has a fivefold increase in bone volume, and a decreased, but more homogeneous, distribution of mineral.


Fig. 5.2 Fourier transform infra-red (FTIR) imaging of the cathepsin K knockout (KO) mouse bones and age- and gender-matched wild-type (WT) controls. (a) A typical spectrum illustrating the peaks of interest from one pixel in the center of WT cortical bone. Parameters calculated for each pixel are mineral/matrix (integrated area of phosphate/amide I bands), crystallinity (peak height ratio of 1030 cm1 /1020 cm1 subbands), and collagen maturity (peak height

ratio of 1660 cm1 /1690 cm1 sub-bands). (b) Hyperspectral images of the mineral/ matrix ratio in growth plate of WT (top; width height ¼ 250 mm 74 mm) and KO (lower; 250 mm 50 mm). (c) Hyperspectral images of crystallinity in same growth plate sections; WT (top) and KO (lower). (d) Mineral/matrix in trabecular bone of WT (top; 2000 mm 1395 mm) and KO (lower; 370 mm 155 mm). (e) Crystallinity in cortical bone of WT (top; 310 mm 250 mm) and KO

In addition to these models, a limited number of examples exist of druginduced osteopetrosis associated with an inhibition of osteoclastic activity. Thus, osteopetrosis has been reported in association with prolonged high dosages of anti-resorptive bisphosphonates (in both rodents [102, 103] and humans [104]), with high prolonged doses of a phytoestrogen in rats [105], and with excessive fluoride consumption [106].

69

70

5 Osteoporosis and Osteopetrosis Table 5.2 Animal models of osteoporosis and osteopetrosis.

Osteoporosis

Reference

Osteopetrosis

Reference(s)

Ovariectomized mouse

22

Spontaneous fracture mouse Steroid-induced rodent Immobilized rodent Senescence-accelerated mouse

29 25 24 27

Naturally occurring rats tl/tl ia/ia

68 69

Ovariectomized monkey Aging monkey Ovariectomized ewe Ovariectomized mini-pig Ovariectomized canine Biglycan KO Osteonectin KO LRP5 KO Sca-1 KO IL-4 transgenic Osteoprotegrin KO Noggin transgenic Sclerostin transgenic Osteogenesis imperfecta mice Mov 13 oim/oim Brtl fro/fro Type I transgenic

58 60 57 59 56 37 33 52 53 48 55 49 51

Naturally occurring mice op/op oc/oc mi/mi gl/gl Cathepsin K KO PU.1 KO c-src KO c-FOS KO Bcl-2 KO Osteocalcin KO RANK Ligand KO Fra-1 transgenic Klotho mutant mice TGF-beta binding protein KO

80 81 82, 83 75, 76 85 93 94 95 96 98 99 97 100 101

40 41 43 44 42

Osteopetrotic cows Avian osteopetrosis OI mouse with excess BP treatment

87–89 90, 91 102, 103

KO ¼ knockout; Klotho ¼ thread of life, an aging model; BP ¼ bisphosphonate.

5.3 The Cellular and Molecular Bases of Osteopetrosis and Osteoporosis 5.3.1 Osteoporosis

Osteoporosis occurs when there is an imbalance of the activities of osteoblasts and osteoclasts. Factors including environment, age, lifestyle, hormonal status, and genetics contribute to this heterogeneous group of diseases. Genetic segregation analysis, family and population studies, and evaluation of congenic mice

5.3 The Cellular and Molecular Bases of Osteopetrosis and Osteoporosis

[107, 108] have identified polymorphisms associated with osteoporosis, including those in the vitamin D receptor gene, a collagen type I gene, the interleukin gene, the Alox gene, and the estrogen receptor alpha gene, among others. Studies of congenic mice have linked candidate genes to bone fragility and high density [108–111] and associated candidate genes with bone fragility [109], but those genes identified in humans account for only @5% of the heritability of osteoporosis [112]. Many of the genes identified are involved in the regulation of the WNT pathway (Fig. 5.3), affecting the activities of both osteoblasts and osteoclasts [113, 114]. As osteoporosis is a complex trait, however, it is unlikely that for most cases a single genetic defect will be identified.

Fig. 5.3 Osteoblast-osteoclast coupling and the role of the Wnt canonical pathway. Cartoon illustrating key factors coupling osteoblast and osteoclast activity. In osteoblasts and pre-osteoblasts, Wnt proteins bind to the transmembrane domain spanning Frizzled receptor (fzR) and LRP5/LRP6 coreceptors. This activates the Dishevelled protein (dsh) by over-phosphorylating it, leading to phosphorylation of b-catenin and degradation of the scaffold complex (GSK-3; glycogen synthase kinase 3), APC (adenomatous polyposis coli), and Axin. Phosphorylation of b-catenin by GSK3 stimulates the degradation of the complex. Stabilized b-catenin accumulates in the

cytosol, translocates to the nucleus (shaded), where it interacts with T-cell factor/lymphoid enhancer binding factor (TCF/LEF) transcription factors to mediate gene transcription, leading to osteoblastogenesis and the inhibition of both osteoblast and osteocyte apoptosis, an increased ratio of osteoprotegrin (OPG) to RANKL, and represses osteoclastogenesis. LRP5/6 coreceptor activity is inhibited by sclerostin (SOST gene product) and (Dkk). Interaction of the Dkk/LRP complex with kremen internalizes the complex for degradation, thus diminishing the number of Wnt coreceptors for signaling.

71

72


A limited number of cases exist where a simple (Mendelian) genetic mutation leads to an altered bone phenotype. For example, a variety of mis-sense mutations leading to single amino acid changes in the amino terminus of the low-density lipoprotein receptor-related protein 5 (LRP5) results in an autosomal dominant high bone mass phenotype [115], without clinical features. Mice in which this gene is knocked out have a severe form of osteoporosis characterized by low bone mass, implying LRP5 might be involved in osteoporosis [116]. Mice lacking both LRP5 and LRP6 have severe limb deformities, and even lower bone mass [117]. A recent large population study of men with osteoporotic fractures has associated variations in both LRP5 and LRP6 [118]. The study of the LRP5/6 mutations is stimulating the analyses of the elements of the WNT-beta catenin pathway for additional osteoporosis-associated genetic defects. The effects of many of these mutated or altered genes have been identified not only from the knockout and transgenic models, but also from studies of osteoblasts and osteoblast progenitor cells derived from these models. As reviewed elsewhere [119], focusing on the factors that regulate the formation of the osteoblast, the abnormalities in these factors, and the coupling of the osteoblast and osteoclast using cell culture systems, provides additional insights into the multitude of factors leading to osteoporosis.

5.3.2 Osteopetrosis

Osteopetrosis is associated with several genes [66, 120]. The genes associated with a human osteopetrosis encode proteins that participate in the functioning of the differentiated osteoclast, while some of the genes identified in animal models affect osteoclastogenesis. Some of the genes that are associated with osteopetrosis based on such studies are listed in Table 5.3. These genes regulate osteoclast number, activity, and/or function. Osteoclastogenesis is dependent on the RANKL system and macrophage-colony stimulating factor [82]. Osteoblasts produce RANK-ligand which activates osteoclast formation and development by binding to a receptor on the osteoclast-forming cells. The osteoblasts also secrete a soluble factor, OPG, that binds to the same receptor and blocks osteoclastogenesis; this is part of the coupling between osteoblasts and osteoclasts. Additionally, matrix proteins present in bone are important for recruiting the osteoclasts to the bone surface. Once on the bone surface, osteoclasts degrade the mineral and organic matrices of bone by secreting hydrochloric acid and proteases (cathepsins, matrix metalloproteases, etc.) [134]. The mineral is dissolved by the hydrochloric acid, and the bicarbonate-chloride channel, its affiliated membrane protein OSM1, and the vacuolar Hþ -adenosine triphosphatase (V-ATPase) are required for removal of the mineral. Matrix metalloproteases, along with the protease cathepsin K, remove the organic matrix. The release of these proteins is important for the recruitment and differentiation of osteoblasts at the bone surface. These osteoblasts then fill in the ‘‘pits’’ made by osteoclasts,

5.3 The Cellular and Molecular Bases of Osteopetrosis and Osteoporosis Table 5.3 The molecular basis of different variants of osteopetrosis.

Gene with mutation

Description

Model

Disease characteristics

Reference(s)

TCIRG1a

a3 -subunit of V-type Hþ -ATPase

Human

Recessive osteopetrosis

121

Atp6Ia

Vacuolar-proton pump, Hþ transporting (member I)

Mouse

Severe osteopetrosis

122

116-kDa V-ATPasea

116 kDa osteoclastspecific vacuolar proton ATPase subunit

Mouse Human

Autosomal recessive lethal

123, 124

Syk

Syk tyrosine kinase

Mouse

Osteopetrotic

125

D11S1889

? – also associated with muscle function

Human

Autosomal dominant – generalized osteosclerosis, most pronounced at the cranial vault – no fractures

126

TRAP5b

Osteoclast-derived serum tartrate-resistant acid phosphatase 5b

Human

Albers–Schonberg disease

127

ClCN7

Chloride channel 7

Human Mouse

Albers–Schonberg disease (autosomal dominant osteopetrosis type II) – large ineffective osteoclasts – gray lethal mouse

127, 129

OSTM1

Osteopetrosis-associated transmembrane protein 1

Human Mouse

Albers–Schonberg disease

128

CAII

Carbonic anhydrase 2

Human

Cranial thickening, distal renal tubular acidosis and increased fractures

130

CSF

Macrophage stimulating factor

Mouse

Deficient macrophages and osteoclasts

131

CATK

Cathepsin K

Mouse Human

Pycnodysostosis

83–85, 132, 133

GL

Gray lethal

Mouse Human

Severe, juvenile, thickened metaphyses

78

a OC116-KDa

(refers also to ATP6i, TCIRGI, a3) subunit of the osteoclast vacuolar Hþ -ATPase (V-Hþ -ATPase) proton pump.

73

74


and consequently there is a ‘‘cycle’’ of formation and resorption, with factors such as parathyroid hormone (PTH), parathyroid hormone-related peptide (PTHrP), prostaglandins, vitamin D, and a variety of growth and transcription factors released by osteoclasts from the bone matrix activating osteoblast precursors to form a new bone matrix, which must then be mineralized.

5.4 Biomineralization in Osteopetrosis and Osteoporosis

An examination of preclinical models of osteopetrosis and osteoporosis reveals that, in general, the mineral (and mechanical) properties are distinct from those in age- and gender-matched controls. The fact that osteopetrosis is predominantly a disease of defective osteoclast activity provokes the thought of why bone formation is impaired. The most likely here answer is that, because osteoblastic activity and osteoclastic activity are coupled, when the osteoclasts cannot remove bone they are unable to release factors that recruit osteoblasts to deposit new bone (collagen and mineral). Thus, there is a decreased activity of both osteoblasts and osteoclasts. The exceptions are those cases with defects in the chloride channel, where tissue culture studies have indicated that isolated osteoblasts or marrow stromal cells can deposit bone in a normal fashion. However, even when osteoblastic activity is decreased, there may be additional bone formation as bone is deposited on calcified cartilage spicules rather than on newly formed osteoid. This results in the presence of smaller mineral crystals and matrices which are rich in cartilage proteins. In osteoporosis, a similar argument holds. It is difficult to determine which comes first: accelerated remodeling in the presence of osteoblasts that cannot match the rate and therefore cannot rebuild bone that has been removed; or decreased formation with osteoclasts that try to remove the new, poorly mineralized matrix before secondary mineralization is complete. In either case there is a vicious cycle, as any new mineralized tissue that forms is rapidly removed, and what is left is an older matrix with larger crystals. The underlying mechanisms for mineral deposition in osteopetrosis and osteoporosis are most likely the same as in the normal individual (as reviewed elsewhere in this volume). The osteoblasts deposit an organic matrix (predominantly collagen), while non-collagenous proteins associated with the collagen provide nucleation sites for the first deposition of mineral crystals, and other proteins regulate the growth of these crystals. Local calcium, phosphate and hydroxide ion concentrations, each regulated by the cell, determine the rates at which mineral can form. In the normal case there is a balance of mineral formation and removal. It is only when these cellular activities become unbalanced – as in osteopetrosis and osteoporosis – that the mechanisms are impaired and the bone properties compromised.

References

Acknowledgments

These studies were supported by NIH grants DE04141, AR037661, AR043125, and AR046121. The author is grateful to Drs. Peter Bullough, Edward DiCarlo, and Joseph M. Lane for their advising and inspiring her to pursue these investigations.

References 1 N.E. Lane, Am. J. Obstet. Gynecol. 2

3 4 5 6

7

8

9 10

11

12

2006, 194, S3–S11. St. Jude’s Hospital Web Site, 2005, http://www.stjude.org/print/ 0,2510,404_2166_2944,00.html. J. Bollerslev, Clin. Genet. 1987, 31, 86–90. J. Bollerslev, P.E. Andersen, Jr., Acta Orthop. Scand. 1989, 60, 110–112. J. Carolino, J.A. Perez, A. Popa, Am. Fam. Physician. 1998, 57, 1293–1296. Y. Gong, M. Vikkula, L. Boon, J. Liu, P. Beighton, R. Ramesar, L. Peltonen, H. Somer, T. Hirose, B. Dallapiccola, A. De Paepe, W. Swoboda, B. Zabel, A. Superti-Furga, B. Steinmann, H.G. Brunner, A. Jans, R.G. Boles, W. Adkins, M.J. van den Boogaard, B.R. Olsen, M.L. Warman, Am. J. Hum. Genet. 1996, 59, 146–151. H. Hartikka, O. Makitie, M. Mannikko, A.S. Doria, A. Daneman, W.G. Cole, L. Ala-Kokko, E.B. Sochett, J. Bone Miner. Res. 2005, 20, 783–789. O. Makitie, A.S. Doria, F. Henriques, W.G. Cole, S. Compeyrot, E. Silverman, R. Laxer, A. Daneman, E.B. Sochett, J. Pediatr. 2005, 146, 395–401. J. Dwyer, Am. J. Clin. Nutr. 2006, 83, 415S–420S. F. Lazner, M. Gowen, D. Pavasovic, I. Kola, Hum. Mol. Genet. 1999, 8, 1839–1846. H.C. Blair, L.J. Robinson, M. Zaidi, Biochem. Biophys. Res. Commun. 2005, 328, 728–738. G. Russell, G. Mueller, C. Shipman, P. Croucher, Novartis Found. Symp. 2001, 232, 251–267.

13 C.G. Keith, Arch. Ophthal. 1968, 79,

234–241. 14 P. Maroteaux, M. Lamy, JAMA 1965,

191, 715–717. 15 A.L. Boskey, in: S.C. Cowin (Ed.),

16 17 18 19 20 21 22 23

24 25

26

27

28

Bone Biomechanics, 3rd edn. CRC Press, Boca Raton, FL, 2001, pp. 5.1– 5.34. M.C. van der Meulen, K.J. Jepsen, B. Mikic, Bone 2001, 29, 101–104. A. Boskey, R. Mendelsohn, J. Biomed. Opt. 2005, 10, 031102–031105. A.L. Boskey, R. Mendelsohn, Vib. Spectrosc. 2005, 38, 107–114. A. Carden, M.D. Morris, J. Biomed. Opt. 2000, 5, 259–268. L.P. Schot, A.H. Schuurs, J. Steroid Biochem. Mol. Biol. 1990, 37, 461–465. D.N. Kalu, Bone Miner. 1991, 15, 175– 191. W.S. Jee, W. Yao, J. Musculoskelet. Neuronal. Interact. 2001, 1, 193–207. C.M. Bagi, S.C. Miller, B.M. Bowman, G.L. Blomstrom, E.P. France, Bone 1992, 13, 35–40. L. Vico, J. Bone Miner. Res. 2003, 18, 561–569. S.C. Manolagas, R.S. Weinstein, J. Bone Miner. Res. 1999, 14, 1061– 1066. N.E. Lane, W. Yao, M. Balooch, R.K. Nalla, G. Balooch, S. Habelitz, J.H. Kinney, L.F. Bonewald, J. Bone Miner. Res. 2006, 21, 466–476. T. Takeda, M. Hosokawa, K. Higuchi, M. Hosono, I. Akiguchi, H. Katoh, Arch. Gerontol. Geriatr. 1994, 19, 185– 192. M.J. Silva, M.D. Brodt, Z. Fan, J.Y. Rho, J. Biomech. 2004, 37, 1639–1646.

75

76

5 Osteoporosis and Osteopetrosis 29 S. Mohan, A. Kapoor, A. Singgih, Z.

30

31

32

33

34

35

36 37

38

39

40

41

42

Zhang, T. Taylor, H. Yu, R.B. Chadwick, Y.S. Chung, L.R. Donahue, C. Rosen, G.C. Crawford, J. Wergedal, D.J. Baylink, J. Bone Miner. Res. 2005, 20, 1597–1610. B. Bar-Shira-Maymon, R. Coleman, A. Cohen, E. Steinhagen-Thiessen, M. Silbermann, Calcif. Tissue Int. 1989, 44, 36–45. M.J. Silva, M.D. Brodt, M. Ko, Y. AbuAmer, J. Bone Miner. Res. 2005, 20, 419–427. M.J. Silva, M.D. Brodt, B. Wopenka, S. Thomopoulos, D. Williams, M.H. Wassen, M. Ko, N. Kusano, R.A. Bank, J. Bone Miner. Res. 2006, 21, 78–88. A.M. Delany, M. Amling, M. Priemel, C. Howe, R. Baron, E. Canalis, J. Clin. Invest. 2000, 105, 915–923. D.T. Gilmour, G.J. Lyon, M.B. Carlton, J.R. Sanes, J.M. Cunningham, J.R. Anderson, B.L. Hogan, M.J. Evans, W.H. Colledge, EMBO J. 1998, 17, 1860–1870. A.L. Boskey, D.J. Moore, M. Amling, E. Canalis, A.M. Delany, J. Bone Miner. Res. 2003, 18, 1005–1011. D. Faibish, S.M. Ott, A.L. Boskey, Clin. Orthop. Relat. Res. 2006, 443, 28–38. T. Xu, P. Bianco, L.W. Fisher, G. Longenecker, E. Smith, S. Goldstein, J. Bonadio, A. Boskey, A.M. Heegaard, B. Sommer, K. Satomura, P. Dominguez, C. Zhao, A.B. Kulkarni, P.G. Robey, M.F. Young, Nat. Genet. 1998, 20, 78–82. A.L. Boskey, M.F. Young, T. Kilts, K. Verdelis, Cells Tissues Organs. 2005, 181, 144–153. S. Viguet-Carrin, P. Garnero, P.D. Delmas, Osteoporosis Int. 2006, 17, 319–336. K.J. Jepsen, M.B. Schaffler, J.L. Kuhn, R.W. Goulet, J. Bonadio, S.A. Goldstein, J. Biomech. 1997, 30, 1141–1147. N.P. Camacho, L. Hou, T.R. Toledano, W.A. Ilg, C.F. Brayton, C.L. Raggio, L. Root, A.L. Boskey, J. Bone Miner. Res. 1999, 14, 264–272. R. Pereira, J.S. Khillan, H.J. Helminen, E.L. Hume, D.J. Prockop, J. Clin. Invest. 1993, 91, 709–716.

43 K.M. Kozloff, A. Carden, C. Bergwitz,

44

45

46

47 48

49 50

51

52

53

54

55

A. Forlino, T.E. Uveges, M.D. Morris, J.C. Marini, S.A. Goldstein, J. Bone Miner. Res. 2004, 19, 614–622. D.O. Sillence, H.E. Ritchie, T. Dibbayawan, D. Eteson, K. Brown, Am. J. Med. Genet. 1993, 45, 276– 283. W. Traub, T. Arad, U. Vetter, S. Weiner, Matrix Biol. 1994, 14, 337– 345. U. Vetter, E.D. Eanes, J.B. Kopp, J.D. Termine, P.G. Robey, Calcif. Tissue Int. 1991, 49, 248–250. J.Y. Reginster, N. Burlet, Bone 2006, 38, S4–S9. D.B. Lewis, H.D. Liggitt, E.L. Effmann, S.T. Motley, S.L. Teitelbaum, K.J. Jepsen, S.A. Goldstein, J. Bonadio, J. Carpenter, R.M. Perlmutter, Proc. Natl. Acad. Sci. USA 1993, 90, 11618–11622. E. Abe, J. Bone Miner. Res. 2000, 15, 663–673. X.B. Wu, Y. Li, A. Schneider, W. Yu, G. Rajendren, J. Iqbal, M. Yamamoto, M. Alam, L.J. Brunet, H.C. Blair, M. Zaidi, E. Abe, J. Clin. Invest. 2003, 112, 924–934. D.G. Winkler, M.K. Sutherland, J.C. Geoghegan, C. Yu, T. Hayes, J.E. Skonier, D. Shpektor, M. Jonas, B.R. Kovacevich, K. Staehling-Hampton, M. Appleby, M.E. Brunkow, J.A. Latham, EMBO J. 2003, 22, 6267– 6276. S.L. Ferrari, S. Deutsch, S.E. Antonarakis, Curr. Opin. Lipidol. 2005, 16, 207–214. M. Bonyadi, S.D. Waldman, D. Liu, J.E. Aubin, M.D. Grynpas, W.L. Stanford, Proc. Natl. Acad. Sci. USA 2003, 100, 5840–5845. N. Bucay, I. Sarosi, C.R. Dunstan, S. Morony, J. Tarpley, C. Capparelli, S. Scully, H.L. Tan, W. Xu, L.D. Lacey, W.J. Boyle, W.S. Simonet, Genes Dev. 1998, 12, 1260–1268. P. Pogoda, M. Priemel, A.F. Schilling, M. Gebauer, P. Catala-Lehnen, F. Barvencik, F.T. Beil, C. Munch, M. Rupprecht, C. Muldner, J.M. Rueger, T. Schinke, M. Amling, J. Bone Miner. Metab. 2005, 23(Suppl.), 97–102.

References 56 M.C. Monier-Faugere, Z. Geng, Q.

57 58

59 60

61

62

63

64

65

66

67 68

69 70

71

72

Qi, I. Arnala, H.H. Malluche, J. Bone Miner. Res. 1996, 11, 446–455. A.S. Turner, Vet. J. 2002, 163, 232– 239. L. Mosekilde, S.E. Weisbrode, J.A. Safron, H.F. Stills, M.L. Jankowsky, D.C. Ebert, C.C. Danielsen, C.H. Sogaard, A.F. Franks, M.L. Stevens, Bone 1993, 14, 379–382. C.P. Jerome, P.E. Peterson, Bone 2001, 29, 1–6. M.D. Grynpas, B. Huckell, K.P. Pritzker, R.G. Hancock, M.J. Kessler, P. R. Health Sci. J. 1989, 8, 197–204. M.D. Grynpas, P. Patterson-Allen, D.J. Simmons, Calcif. Tissue Int. 1986, 39, 57–62. K. Lundon, M. Dumitriu, M. Grynpas, Bone Miner. 1994, 24, 135– 149. S.J. Gadeleta, A.L. Boskey, E. Paschalis, C. Carlson, F. Menschik, T. Baldini, M. Peterson, C.M. Rimnac, Bone 2000, 27, 541–550. D. Pienkowski, T.M. Doers, M.C. Monier-Faugere, Z. Geng, N.P. Camacho, A.L. Boskey, H.H. Malluche, J. Bone Miner. Res. 1997, 12, 1936–1943. L.M. Miller, J. Tibrewala, C.S. Carlson, Cell Mol. Biol. (Noisy-legrand) 2000, 46, 1035–1044. W. Balemans, L. Van Wesenbeeck, W. Van Hul, Calcif. Tissue Int. 2005, 77, 263–274. S.C. Marks, Jr., Clin. Orthop. Relat. Res. 1984, 189, 239–263. A. Wojtowicz, A. Dziedzic-Goclawska, A. Kaminski, W. Stachowicz, K. Wojtowicz, S.C. Marks, Jr., M. Yamauchi, Bone 1997, 20, 127–132. A.L. Boskey, S.C. Marks, Jr., Calcif Tissue Int. 1985, 37, 287–292. N. Sakagami, N. Amizuka, M. Li, K. Takeuchi, M. Hoshino, M. Nakamura, K. Nozawa-Inoue, N. Udagawa, T. Maeda, Micron 2005, 36, 688–695. M.E. Jackson, V. Shalhoub, J.B. Lian, G.S. Stein, S.C. Marks, Jr., J. Cell Biochem. 1994, 55, 366–372. T.J. Martin, N.A. Sims, Trends Mol. Med. 2005, 11, 76–81.

73 J. Tuukkanen, A. Koivukangas, T.

74 75

76

77

78

79

80

81

82

83

84

85

86

Jamsa, K. Sundquist, C.A. Mackay, S.C. Marks, Jr., J. Bone Miner. Res. 2000, 15, 1905–1911. A. Boskey, Crit. Rev. Eukaryot. Gene Expr. 2003, 13, 109–116. U. Kornak, D. Kasper, M.R. Bo¨sl, E. Kaiser, M. Schweizer, A. Schulz, W. Friedrich, G. Delling, T.J. Jentsch, Cell 2001, 104, 205–215. N. Chalhoub, N. Benachenhou, V. Rajapurohitam, M. Pata, M. Ferron, A. Frattini, A. Villa, J. Vacher, Nat. Med. 2003, 9, 399–406. A. Ramirez, J. Faupel, I. Goebel, A. Stiller, S. Beyer, C. Stockle, C. Hasan, U. Bode, U. Kornak, C. Kubisch, Hum. Mutat. 2004, 23, 471–476. P. Quarello, M. Forni, L. Barberis, C. Defilippi, M.F. Campagnoli, L. Silvestro, A. Frattini, N. Chalhoub, J. Vacher, U. Ramenghi, J. Bone Miner. Res. 2004, 19, 1194–1199. S. Schaller, K. Henriksen, M.G. Sorensen, M.A. Karsdal, Drug News Perspect. 2005, 18, 489–495. J.C. Scimeca, A. Franchi, C. Trojani, H. Parrinello, J. Grosgeorge, C. Robert, O. Jaillon, C. Poirier, P. Gaudray, G.F. Carle, Bone 2000, 26, 207–213. K. Roundy, R. Smith, J.J. Weis, J.H. Weis, J. Bone Miner. Res. 2003, 18, 278–288. T. Wada, T. Nakashima, N. Hiroshi, J.M. Penninger, Trends Mol. Med. 2006, 12, 17–25. W.S. Hou, D. Bromme, Y. Zhao, E. Mehler, C. Dushey, H. Weinstein, C.S. Miranda, C. Fraga, F. Greig, J. Carey, D.L. Rimoin, R.J. Desnick, B.D. Gelb, J. Clin. Invest. 1999, 103, 731–738. N. Fratzl-Zelman, A. Valenta, P. Roschger, A. Nader, B.D. Gelb, P. Fratzl, K. Klaushofer, J. Clin. Endocrinol. Metab. 2004, 89, 1538–1547. M. Gowen, F. Lazner, R. Dodds, R. Kapadia, J. Feild, M. Tavaria, I. Bertoncello, F. Drake, S. Zavarselk, I. Tellis, P. Hertzog, C. Debouck, I. Kola, J. Bone Miner. Res. 1999, 14, 1654–1663. C.Y. Li, K.J. Jepsen, R.J. Majeska, J. Zhang, R. Ni, B.D. Gelb, M.B.

77

78


87 88

89 90 91 92

93

94 95 96

97

98

99

100

101

Schaffler, J. Bone Miner. Res. 2006, 21, 865–875. H.W. Leipold, J.E. Cook, Am. J. Pathol. 1977, 86, 745–748. R.B. Ashman, W.C. Van Buskirk, S.C. Cowin, P.M. Sandborn, M.K. Wells, J.C. Rice, Calcif. Tissue Int. 1985, 37, 73–76. D.C. Wolf, W.G. Van Alstine, J. Vet. Diagn. Invest. 1989, 1, 262–264. E.V. Schmidt, R.E. Smith, Am. J. Pathol. 1982, 106, 297–299. R.E. Smith, Curr. Top Microbiol. Immunol. 1982, 101, 75–94. R. Michael McClain, E. Wolz, A. Davidovich, F. Pfannkuch, J.A. Edwards, J. Bausch, Food Chem. Toxicol. 2006, 44, 56–80. M.M. Tondravi, S.R. McKercher, K. Anderson, J.M. Erdmann, M. Quiroz, R. Maki, S.L. Teitelbaum, Nature 1997, 386, 81–84. P. Soriano, C. Montgomery, R. Geske, A. Bradley, Cell 1991, 64, 693–702. O. Jacenko, BioEssays 1995, 17, 277– 281. G.G. McGill, M. Horstmann, H.R. Widlund, J. Du, G. Motyckova, E.K. Nishimura, Y.L. Lin, S. Ramaswamy, W. Avery, H.F. Ding, S.A. Jordan, I.J. Jackson, S.J. Korsmeyer, T.R. Golub, D.E. Fisher, Cell 2002, 109, 707–718. P. Roschger, K. Matsuo, B.M. Misof, W. Tesch, W. Jochum, E.F. Wagner, P. Fratzl, K. Klaushofer, Bone 2004, 34, 776–782. A.L. Boskey, S. Gadaleta, C. Gundberg, S.B. Doty, P. Ducy, G. Karsenty, Bone 1998, 23, 187–196. Y.Y. Kong, H. Yoshida, I. Sarosi, H.L. Tan, E. Timms, C. Capparelli, S. Morony, A.J. Oliveira-dos-Santos, G. Van, A. Itie, W. Khoo, A. Wakeham, C.R. Dunstan, D.L. Lacey, T.W. Mak, W.J. Boyle, J.M. Penninger, Nature 1999, 397, 315–323. T. Yamashita, I. Sekiya, N. Kawaguchi, K. Kashimada, A. Nifuji, Y.I. Nabeshima, M. Noda, J. Endocrinol. 2001, 168, 347–351. B. Dabovic, R. Levasseur, L. Zambuto, Y. Chen, G. Karsenty, D.B. Rifkin, Bone 2005, 37, 25–31.

102 R. Schenk, W.A. Merz, R. Muhlbauer,

103

104

105

106 107

108

109

110

111

112

113

114

115

R.G. Russell, H. Fleisch, Calcif. Tissue Res. 1973, 11, 196–214. N.P. Camacho, C.L. Raggio, S.B. Doty, L. Root, V. Zraick, W.A. Ilg, T.R. Toledano, A.L. Boskey, Calcif. Tissue Int. 2001, 69, 94–101. M.P. Whyte, D. Wenkert, K.L. Clements, W.H. McAlister, S. Mumm, N. Engl. J. Med. 2003, 349, 457–463. R. Michael McClain, E. Wolz, A. Davidovich, F. Pfannkuch, J.A. Edwards, J. Bausch, Food Chem. Toxicol. 2006, 44, 56–80. M. Grynpas, Calcif. Tissue Int. 1993, Suppl. 1, S57–S64. Q.Y. Huang, R.R. Recker, H.W. Deng, Osteoporos. Int. 2003, 14, 701–715. R.F. Klein, J. Allard, Z. Avnur, T. Nikolcheva, D. Rotstein, A.S. Carlos, M. Shea, R.V. Waters, J.K. Belknap, G. Peltz, E.S. Orwoll, Science 2004, 303, 229–232. Y. Yershov, T.H. Baldini, S. Villagomez, T. Young, M.L. Martin, R.S. Bockman, M.G. Peterson, R.D. Blank, J. Bone Miner. Res. 2001, 16, 992–1003. M. Shimizu, K. Higuchi, S. Kasai, T. Tsuboyama, M. Matsushita, T. Matsumura, S. Okudaira, M. Mori, A. Koizumi, T. Nakamura, M. Hosokawa, Mamm. Genome 2002, 13, 335–340. W. Gu, X. Li, K.H. Lau, B. Edderkaoui, L.R. Donahae, C.J. Rosen, W.G. Beamer, K.L. Shultz, A. Srivastava, S. Mohan, D.J. Baylink, Funct. Integr. Genomics 2002, 1, 375– 386. F.M. Williams, T.D. Spector, J. Musculoskelet. Neuronal Interact. 2006, 6, 27–35. D.A. Glass, II, P. Bialek, J.D. Ahn, M. Starbuck, M.S. Patel, H. Clevers, M.M. Taketo, F. Long, A.P. McMahon, R.A. Lang, G. Karsenty, Dev. Cell 2005, 8, 751–764. V. Krishnan, H.U. Bryant, O.A. Macdougald, J. Clin. Invest. 2006, 116, 1202–1209. L. Van Wesenbeeck, E. Cleiren, J. Gram, R.K. Beals, O. Benichou, D.

References

116

117

118

119 120

121

122

123

Scopelliti, L. Key, T. Renton, C. Bartels, Y. Gong, M.L. Warman, M.C. De Vernejoul, J. Bollerslev, W. Van Hul, Am. J. Hum. Genet. 2003, 72, 763–771. M. Kato, M.S. Patel, R. Levasseur, I. Lobov, B.H. Chang, D.A. Glass, II, C. Hartmann, L. Li, T.H. Hwang, C.F. Brayton, R.A. Lang, G. Karsenty, L. Chan, J. Cell Biol. 2002, 157, 303–314. S.L. Holmen, T.A. Giambernardi, C.R. Zylstra, B.D. Buckner-Berghuis, J.H. Resau, J.F. Hess, V. Glatt, M.L. Bouxsein, M. Ai, M.L. Warman, B.O. Williams, J. Bone Miner. Res. 2004, 19, 2033–2040. J.B. van Meurs, F. Rivadeneira, M. Jhamai, W. Hugens, A. Hofman, J.P. van Leeuwen, H.A. Pols, U.A. Gitterlinden, J. Bone Miner. Res. 2006, 21, 141–150. D. Rowe, A. Lichtler, Endocrine 2002, 17, 67–75. C. Sobacchi, A. Frattini, P. Orchard, O. Porras, I. Tezcan, M. Andolina, R. Babul-Hirji, I. Baric, N. Canham, D. Chitayat, S. Dupuis-Girod, I. Ellis, A. Etzioni, A. Fasth, A. Fisher, B. Gerritsen, V. Gulino, E. Horwitz, V. Klamroth, E. Lanino, M. Mirolo, A. Musio, G. Matthijs, S. Nonomaya, L.D. Notarangelo, H.D. Ochs, A. Superti Furga, J. Valiaho, J.L. van Hove, M. Vihinen, D. Vujic, P. Vezzoni, A. Villa, Hum. Mol. Genet. 2001, 10, 1767–1777. A. Frattini, P.J. Orchard, C. Sobacchi, S. Giliani, M. Abinun, J.P. Mattsson, D.J. Keeling, A.K. Andersson, P. Wallbrandt, L. Zecca, L.D. Notarangelo, P. Vezzoni, A. Villa, Nat. Genet. 2000, 25, 343–346. Y.P. Li, W. Chen, Y. Liang, E. Li, P. Stashenko, Nat. Genet. 1999, 23, 447– 451. J.C. Scimeca, A. Franchi, C. Trojani, H. Parrinello, J. Grosgeorge, C. Robert, O. Jaillon, C. Poirier, P.

124 125

126

127

128

129

130

131

132 133

134

Gaudray, G.F. Carle, Bone 2000, 26, 207–213. K.U. Ogbureke, Q. Zhao, Y.P. Li, Front. Biosci. 2005, 10, 2940–2954. A. Mocsai, M.B. Humphrey, J.A. Van Ziffle, Y. Hu, A. Burghardt, S.C. Spusta, S. Majumdar, L.L. Lanier, C.A. Lowell, M.C. Nakamura, Proc. Natl. Acad. Sci. USA 2004, 101, 6158– 6163. E. Van Hul, J. Gram, J. Bollerslev, L. Van Wesenbeeck, D. Mathysen, P.E. Andersen, F. Vanhoenacker, W. Van Hul, J. Bone Miner. Res. 2002, 17, 1111–1117. S.L. Alatalo, K.K. Ivaska, S.G. Waguespack, M.J. Econs, H.K. Vaananen, J.M. Halleen, Clin. Chem. 2004, 50, 883–890. P.F. Lange, L. Wartosch, T.J. Jentsch, J.C. Fuhrmann, Nature 2006, 440, 220–223. U. Kornak, A. Ostertag, S. Branger, O. Benichou, M.C. de Vernejoul, J. Clin. Endocrinol. Metab. 2006, 91, 995–1000. K.J. Borthwick, N. Kandemir, R. Topaloglu, U. Kornak, A. Bakkaloglu, N. Yordam, S. Ozen, H. Mocan, G.N. Shah, W.S. Sly, F.E. Karet, J. Med. Genet. 2003, 40, 115–121. W. Wiktor-Jedrzejczak, A. Bartocci, A.W. Ferrante, Jr., A. Ahmed-Ansari, K.W. Sell, J.W. Pollard, E.R. Stanley, Proc. Natl. Acad. Sci. USA 1990, 87, 4828–4832. G. Motyckova, D.E. Fisher, Curr. Mol. Med. 2002, 2, 407–421. P. Saftig, E. Hunziker, O. Wehmeyer, S. Jones, A. Boyde, W. Rommerskirch, J.D. Moritz, P. Schu, K. von Figura, Proc. Natl. Acad. Sci. USA 1998, 95, 13453–13458. K. Henriksen, M.G. Sorensen, R.H. Nielsen, J. Gram, S. Schaller, M.H. Dziegiel, V. Everts, J. Bollerslev, M.A. Karsdal, J. Bone Miner. Res. 2006, 21, 58–66.

79

81

6 Biomimetic Bone Substitution Materials Matthias Epple

Abstract

Bone defects are filled with different materials, with the patient’s own bone (e.g., harvested from the iliac crest) being the ‘‘gold standard’’ in surgery. However, the insufficient supply of this autologous bone in the case of major defects, and the need for a second operation to harvest this bone, have triggered research into semi- or fully synthetic bone substitution materials. In general, it is desirable that a bone defect is filled by newly grown bone after some time; that is, the implant material should be biodegradable and permit or (even better) stimulate the ingrowth of bone. Thus, the regeneration depends on the body’s own restorative capability, and it may be assumed that a material which has properties close to natural bone will be advantageous. Within the current concepts for bone substitution materials, the role of biomimetic – that is, ‘‘bone-resembling’’ – implants is highlighted. Keywords: implants.

surgery, bone, calcium phosphate, polymers, metals, osteoblasts,

6.1 The Clinical Need for Bone Substitution Materials

There are a number of clinical situations where lost bone has to be replaced. Typical examples are complicated fractures, explantation sites of bone tumors, bone loss around endoprostheses, and bone loss in the jaw around lost or extracted teeth [1–5]. In all of these cases, the bone defect must be filled with a suitable material which has a sufficient mechanical stability, does not cause chemically adverse reactions (e.g., the release of acids or toxic metals), and does not lead to an adverse biological reaction (e.g., inflammation or an allergic reaction). Ideally, the material should be biodegradable and eventually be replaced by newly grown


82

6 Biomimetic Bone Substitution Materials

bone, leading to a restitutio ad integrum of the bone defect. Note that bone as a living tissue is undergoing a constant biomineralization, the so-called ‘‘remodeling’’; that is, osteoclasts are continuously resorbing bone and osteoblasts are simultaneously forming new bone [6]. Therefore, the process of the replacement of an implant by new bone must rely on the same processes as the biomineralization in healthy bone. Consequently, the most straightforward approach is the use of explanted native bone from a donor site of the same patient, and its re-implantation into the defect site. Usually, such a transplant is well accepted by the body and rapidly integrated into the surrounding bone tissue. Therefore, this autologous bone implant is usually referred to as the ‘‘gold standard’’ in clinical medicine. However, the fact that there is not much spare bone in the human body (typically, bone is harvested from the iliac crest), and that there is sometimes a rapid resorption of the implanted bone (faster than the growth of new bone), sets a limit to this approach. The need for an additional operation to explant the bone from the donor site, which often is accompanied by additional pain for the patient, is another restriction. Due to high clinical demand, many different semi- or fully synthetic materials have been proposed to treat bone defects. Bone from other donors (from bone banks, so-called ‘‘allogenic transplants’’) can also be implanted, but this is limited by the need to suppress an adverse immune reaction and to exclude transmitted infections. Nevertheless, such allogenic transplants still constitute a considerable part of clinical practice, due also to the fact that there is often ‘‘spare bone’’ available, for example from the removed femoral heads after the implantation of an artificial hip joint. The next logical step is to implant bone from animals – the so-called ‘‘xenogenic transplants’’ – where there is a potentially unlimited supply. However, concerns about immune reactions and infections are even more severe in this case, and therefore xenogenic bone can only be used after extensive chemical and/or thermal treatment in order to exclude all hazardous biological material. Thus, fully synthetic biomaterials for bone substitution offer great potential, provided that they fulfill all mechanical, chemical and biological requirements. The high economic opportunities associated with this process have resulted in many different biomaterials for this application.

6.2 Synthetic Materials Used for Bone Substitution

Materials science is a well-developed field of science, and in recent years many materials have been proposed as bone substitution materials, with a more or less biological relationship to the original material, living bone. The main requirements for a bone substitute include:

6.2 Synthetic Materials Used for Bone Substitution

A sufficient mechanical stability which, ideally, is identical to that of bone. A low mechanical stability leads to disintegration and undesired destabilization of the implantation site. Conversely, a high mechanical stability, characterized by a high stiffness (high module of elasticity), leads to stress-shielding of the surrounding bone and potentially to bone loss around the implant. A biodegradability which is adapted to the biological requirements; that is, it should be rapid enough to allow new bone to grow into the implantation site, but not so rapid that a mechanically weak point results. Ideally, the combined mechanical strength of the implant and of the ingrowing bone should remain constant throughout the regenerative process. A high porosity, which allows the ingrowth of bone tissue during regeneration. This requires typical pore sizes of a few hundred micrometers, which were shown to be well-suited to cell invasion [7]. For a good bone ingrowth the pores should be interconnected, as in native bone (i.e., not isolated). The absence of any chemical or biological irritation by substances which are released due to corrosion or degradation. Typical chemically adverse reactions are the release of immunogenic metal ions (e.g., nickel) and the release of lactic acid during the degradation of poly(lactic acid). An absence of the release of biologically adverse substances, such as immunogenic (e.g., proteins), infectious (e.g., viruses, bacteria), or toxic compounds. A possibility to adjust the shape of the implant during the operation in order to fulfill the surgeons’ requirements. Another possibility is a pre-shaping of the implant before surgery, based on a previous geometric analysis of the bone defect (e.g., by microtomography). A good sterilizability, storability, and processability. A price which is low enough to permit a clinical application.

These requirements are clearly manifold, and impossible to fulfill with any single material. This may explain why autologous bone is still the ‘‘gold standard’’, and why synthetic approaches try to mimic the original bone as closely as possible. Some examples of fully synthetic commercial bone substitution materials are shown in Figure 6.1.

83

84


Fig. 6.1 Some commercial synthetic bone substitution materials. (A) A granulate of hydroxyapatite, Ca10 (PO4 )6 (OH)2 which is used to fill bone defects, sometimes after mixing with blood or bone chips. (B) A bone cement consisting of a solid mixture of calcium phosphates in a small elastic balloon, an aqueous NaCl solution, and a syringe to mix these components (from left to right). The NaCl solution is added to the calcium phosphate powder in the balloon and temporary dissolution, followed by precipitation of calcium phosphate, occurs. This dispersion can be introduced into a bone defect where it hardens in situ.

(C) A granulate of b-tricalcium phosphate (b-Ca3 (PO4 )2 ; b-TCP) for the same application as in case (A). (D) Cylinders of nickeltitanium (NiTi), which are used as nonbiodegradable, weight-bearing implants in the spine. (E) A paste of nano-hydroxyapatite in water with high mineral content which can be injected into bone defects. Note that hardening does not occur in this case because no precipitation occurs, and because the paste will not dry in the ‘‘moist’’ biological environment. (F) Porous blocks of b-TCP which are individually machined and can be used as three-dimensional defect-filling materials.

6.3 Ceramics and Bone Cements

Bone and dentin mineral consists of nanocrystals of carbonated apatite, the formula of which may expressed simplified as Ca10-x (PO4 )6-x (CO3 )x (OH)2 , and sometimes denoted as dahllite [8, 9]. Interestingly, the same mineral in terms of composition and crystal size was found in atherosclerotic lesions [10], pointing to similar formation pathways. A whole range of different calcium phosphates has been identified [9, 11–13], and in general all are biocompatible due to their similarity to bone and tooth mineral. Consequently, many attempts have been un-

6.3 Ceramics and Bone Cements

dertaken to use calcium phosphates as bone substitution materials, and today many different products are available internationally, though most are based on hydroxyapatite (HAP or HA), Ca10 (PO4 )6 (OH)2 , or b-tricalcium phosphate (bTCP), or b-Ca3 (PO4 )2 [14]. Some current problems involve the inherent brittleness of ceramics, as this may lead to mechanical failure at the operation site and a sometimes inadequate biodegradation (often too slow). The first of these problems can be solved by using sintering processes which increase the hardness of a material; this in turn often slows down the rate of biodegradation [15–18]. The sometimes slow rate of biodegradation can be understood when the biological mechanism for degradation of calcium phosphates is considered. Bone tissue is continually undergoing a permanent remodeling process [19]; that is, old bone is being resorbed by osteoclasts and new bone is being formed by osteoblasts [5, 20]. Osteoclasts are also responsible for the biodegradation of calcium phosphate implants [18, 21], and function by creating a secluded compartment between the cell and the bone, characterized by the so-called ‘‘ruffled border’’ of the osteoclast. An acidic pH of about 4, caused by the presence of hydrochloric acid, is created by proton and chloride pumps [22–25], and this in turn leads to a dissolution of the nanoscopic bone mineral crystals, because all calcium phosphates are soluble in acids [9, 26]. Because calcined ceramics consist of microcrystals instead of nanocrystals, they have a lower solubility and are dissolved only slowly under the conditions of osteoclastic resorption (compare the morphology of the material in Figs. 6.3 and 6.4 at high magnification). This was also shown experimentally in vitro [16, 27–32]. It is therefore important to assure a good solubility of the calcium phosphate under the conditions of osteoclastic dissolution [18], either by choosing a phase with higher solubility (e.g., b-TCP [33], octacalcium phosphate, OCP [34], or carbonated apatite [22, 35–38]), or by keeping the size of the crystals within the nanometer range [38–42]. Slowly degrading ceramic implants may cause problems if further traumatic fractures occur at the same site [43]. Bone cements consisting of carbonated apatite can be precipitated in situ by mixing powders and solutions which contain the components of carbonated apatite, and injecting the resulting paste into the bone defect [44–46]. Such bone cements have gained some clinical acceptance and show a good biodegradability [31]. Glass ceramics (‘‘Bioglasses’’), which are based on ‘‘CaOP2 O5 SiO2 ’’ are also used as bone substitution materials, with an adjustable range of properties, depending on the composition [47, 48]. In summary, it appears to be advantageous to come as close as possible to the biological example – that is, to nanosized carbonated apatite. Unfortunately, although it is possible to prepare bone mineral-like calcium phosphate nanoparticles [40, 49, 50], it has not yet been possible to mimic completely the finely structured composite of collagen-calcium phosphate by synthetic means, despite promising attempts [51–58]. After all, this hierarchical structure is responsible for the exceptional mechanical properties of bone ! [6, 59, 60]. Although calcium phosphates are by far the most important ceramics to be used as bone substitution materials, it has also been proposed to use calcium sulfate [61] or calcium carbonate [62, 63] in such a role. Unfortunately, however,

85

86


both materials show a rather rapid biodegradation, due either to their high solubility either in general (calcium sulfate) or under acidic conditions (calcium carbonate).

6.4 Polymers

Polymers have a greater elasticity than ceramics, and furthermore their properties can be fine-tuned over a wide range by varying the type of polymer, its chain length and crystallinity, and/or by preparing co-polymers and by blending two or more polymers [64–66]. In particular, this permits the adjustment of mechanical properties and of the rate of biodegradation [67]. If biodegradable [68] polymers are to be used, it must be ensured that the degradation products – that is, the monomers and oligomers of the polymer – do not induce any adverse reactions. Consequently, the number of polymers in clinical use is limited. For bone substitution, two main synthetic classes of materials are used: (i) polymethylmethacrylate (PMMA) and its derivatives; and (ii) various polyesters from hydroxycarboxylic acids [67]. 6.4.1 PMMA-Based Materials

Bone cements based on PMMA are applied as a paste of the liquid monomer and solid particles of oligomers, together with an initiator [69, 70]. The material polymerizes on the implantation site and integrates as a tough, very stable implant. It is often applied to fix total hip endoprostheses in the femoral leg. The drawbacks of PMMA use are the heat evolution during polymerization, which may lead to necrosis of the surrounding bone, and the release of small amounts of oligomers into the surrounding tissue. PMMA is not biodegradable and does not induce the ongrowth of bone. 6.4.2 Polyester-Based Materials

Polyesters such as polyglycolide (polyglycolic acid; PGA) and polylactide (polylactic acid; PLA) and co-polyesters of these, are applied in medicine as biodegradable implant materials [67, 71–74]; they have also been studied extensively as bone substitution materials [75–80] and as scaffolds for tissue engineering [81–85]. They show good mechanical characteristics (they are more elastic than ceramics) and degrade to the corresponding hydroxycarboxylic acids, which are easily metabolized [71]. Occasionally, an accumulation of the acidic degradation products has been observed which led to serious inflammation and damage of the surrounding tissues [86, 87]. However, this can be countered by adding basic salts such as calcium

6.7 Bone Substitutes of Biological Origin

carbonate or carbonated apatite to the materials [88–91]; this has the added advantage of making the material more biocompatible.

6.5 Metals

Metals are often used in medicine as surgical materials, mainly as plates, nails or screws. For bone substitution, very few (and usually porous [92]) materials have been proposed, including porous titanium [93], porous nickel-titanium (NiTi) [94–97], porous tantalum [98], and magnesium alloys [99]. Of course, they are all not biomimetic in terms of chemical composition because there are no elemental metals in the human body, but they can show a good biocompatibility in bone contact, especially when they are coated with calcium phosphate (the bone mineral [100–106]; see also Chapter 7). Usually, these are permanent implants in which degradation or corrosion is not desirable, although during recent years a number of magnesium alloys have been proposed which are aimed to degrade in the body in order to make room for the ingrowing bone [99].

6.6 Composites

Bone is a composite material which consists mainly of collagen (‘‘an elastic polymer’’) and calcium phosphate (‘‘a tough ceramic’’). Its extraordinary mechanical properties of being simultaneously elastic and hard [6, 107] suggest that just one replacement material alone will not be able to fulfill all requirements. Consequently, investigations are ongoing in order to develop composite materials from polymers and ceramics which are aimed to mimic the properties of natural bone. Many different polymers have been combined with calcium phosphate as ceramic filler material; calcium phosphate is usually chosen because of its excellent biocompatibility, its biodegradability, and its good mechanical properties when used as a filler material. For a review on polymer/calcium phosphate composite materials used for bone substitution, the reader is referred to Ref. [108].

6.7 Bone Substitutes of Biological Origin

Nacre is well received by the body upon implantation, and may even contain osteoinductive substances [62, 109, 110]. Other biological materials comprise corals [111–13], chemically or thermally treated bone xenografts [114], and hydrothermally treated calcareous algae [14, 115]. The latter material consists of a calcium carbonate skeleton which is converted into hydroxyapatite or b-tricalcium phosphate by hydrothermal treatment with ammonium phosphate [115]. The external

87

88

6 Biomimetic Bone Substitution Materials Fig. 6.2 A bone substitution material which was obtained by hydrothermal conversion of the calcareous skeleton of red algae into hydroxyapatite. Note the preservation of the porous skeleton of the algae in the final product.

shape and internal porosity are well preserved (Fig. 6.2). This material demonstrated a good clinical success, especially in oral surgery, and its performance can be improved my mixing with autologous bone chips [116, 117]. Juniper wood has also been proposed as a bone substitution material [118]. It is important to note that all materials of biological origin must be suitably conditioned before any implantation due to the risk of infection, and this may involve chemical and/or thermal treatment. A bone substitution material which was derived from bovine bone by chemical and thermal treatment is shown in Figure 6.3 [14]. Of particular note is the porous structure, as well as the non-resolved microstructure at the highest magnification, which shows no individual crystals and a morphology which is close to that of original bone. The sample still consists of collagen and calcium phosphate nanocrystals. A bone substitution material obtained by the calcination of bovine bone is shown in Figure 6.4 [14]. The interconnecting pore structure of the original bone is still present, but the former nanocrystals have sintered into microcrystals (these are clearly visible at the highest magnification). The driving forces for these developments were both the chemical similarity to bone (mineral) and the morphological similarity to cancellous bone, which allows an easy ingrowth of bone. The graded nature of bone (cortical and cancellous bone) is also an important property to reproduce in bone substitution materials, not only to control the

6.7 Bone Substitutes of Biological Origin Fig. 6.3 A bone substitution material which was obtained by thermal and chemical treatment of bovine bone (to remove all infectious components). Chemically, the sample consists of collagen and nanocrystalline calcium phosphate. Morphologically, the interconnecting porosity of the original bone is preserved.

Fig. 6.4 A bone substitution material which was obtained by calcination of bovine bone. It consists exclusively of inorganic components, mainly hydroxyapatite from the original bone mineral, and preserves the interconnecting porosity of natural bone. On the microscale, however, it is clear that the calcium phosphate nanocrystals have sintered into microcrystals.

89

90


mechanical properties but also to achieve spatially different rates of biodegradation [119, 120].

6.8 Biological Functionalization of Synthetic Materials

One concept which goes beyond the pure chemical, structural or morphological variation of a synthetic biomaterial involves its functionalization with biologically active molecules. These can be either substances which enhance bone growth after being released (e.g., morphogenetic proteins [121–123]; see also Chapter 2), or substances which attract cells to the implant surface, making use of special recognition sequences such as integrin ligands ([124, 125]; see also Chapter 8). These approaches must be distinguished on a conceptual basis, however: the first method induces bone formation due to drug release in the vicinity of the implant, whereas the second method modifies the surface only – that is, it acts in a spatially defined manner. Synthetic bone substitution materials can be used as carriers for biomolecules, with bone growth being induced in their vicinity during release (see, e.g., Refs. [112, 116, 126–130]).

6.9 An Example of a Synthetic Biomimetic Bone Substitution Material

The general concepts behind the preparation of biomimetic materials involve biodegradable materials, materials with a ‘‘bone-like’’ porous structure, and a mechanical performance resembling that of bone. If a biological functionalization is set aside, this must be achieved using synthetic or biologically derived materials. Such an example is described in the following section. A combined effort made by chemists, engineers, and clinicians led to an individually shaped skull implant which consisted of polylactides and calcium phosphate/calcium carbonate (fully biodegradable), and was structured according to the biological requirements at the implantation site [79, 80, 131]. The implant was designed to mimic the cancellous and cortical structure of bone; that is, it is a functionally graded material like natural bone [6] (Fig. 6.5). The porous inside (‘‘cancellous bone’’) consisted of rapidly degrading poly(dl-lactide) and calcium carbonate, pointed towards the dura mater, and permitted the ingrowth of bone from that region. The compact exterior (‘‘cortical bone’’), consisted of slowly degrading poly(l-lactide) and nanoscopic carbonated apatite. An integrated micro computed tomography/computer-aided manufacturing process chain (‘‘TICC’’; [79, 80]) permitted a patient-specific shaping of the implant. Subsequent experiments conducted in animals showed the implant to have excellent biocompatibility and to be almost completely degraded and/or substituted by newly grown bone [80].

Acknowledgments

Fig. 6.5 A biomimetic bone substitution implant for the treatment of skull defects. The part pointing towards the brain (top) consists of porous poly(dl-lactide), together with calcium carbonate, the part pointing towards the skin (bottom) consists of compact poly(l-lactide), together with nano-calcium phosphate.

6.10 Conclusions and Future Developments

While the in-vitro preparation of new bone by tissue engineering remains in its infancy, despite considerable efforts having been made (see Chapter 10), surgeons will continue to rely either on autologous bone grafts or on synthetic biomaterials. This will involve strategies that are known from Nature, and include materials which are biodegradable, mechanically and morphologically optimized, and possibly also nano-structured. Although this can be achieved by using synthetic materials, in future the biological functionalization of implants will attract an increasing amount of attention. This functionalization will be introduced not only onto the implant surface in order to achieve interaction with the surrounding tissue and cells, but also internally, perhaps to incorporate drugs and biomolecules that will promote healing and induce future bone growth. Clearly, these modifications will involve more costly preparations and formulations than are presently in use, and competition will no doubt evolve between the design and development of biologically optimized implants and cost-limiting procedures of the social security systems.

Acknowledgments

The biomimetic skull implant was developed in a major project funded by the Deutsche Forschungsgemeinschaft (DFG). The author wishes to acknowledge his partners and co-workers, including Thomas Annen, Harald Eufinger, Chris-

91

92


tian Rasche, Carsten Schiller, Inge Schmitz, Michael Wehmo¨ller, and Stephan Weihe, all of whom have contributed to these results.

References 1 R.A. Kenley, K. Yim, J. Abrams, E.

2 3

4 5

6 7

8 9

10

11

12

13 14 15

16

Ron, T. Turek, L.J. Marden, J.O. Hollinger, Pharmaceut. Res. 1993, 10, 1393–1401. J.M. Rueger, Orthopa¨de 1998, 27, 72– 79. A.S. Greenwald, S.D. Boden, V.M. Goldberg, Y. Khan, C.T. Laurencin, R.N. Rosier, J. Bone Joint Surg. 2001, 83A, 98–103. S.V. Dorozhkin, M. Epple, Angew. Chem. 2002, 114, 3260–3277. M. Amling, A.F. Schilling, P. Pogoda, M. Priemel, J.M. Rueger, Eur. J. Trauma 2006, 32, 102–106. S. Weiner, H.D. Wagner, Annu. Rev. Mater. Sci. 1998, 28, 271–298. S.L. Ishaug, G.M. Crane, M.J. Miller, A.W. Yasko, M.J. Yaszemski, A.G. Mikos, J. Biomed. Mater. Res 1997, 36, 17–28. U. Akiva, H.D. Wagner, S. Weiner, J. Mater. Sci. 1998, 33, 1497–1509. S.V. Dorozhkin, M. Epple, Angew. Chem. Int. Ed. 2002, 41, 3130– 3146. A. Becker, M. Epple, K.M. Mu¨ller, I. Schmitz, J. Biol. Inorg. Chem. 2004, 98, 2032–2038. R.Z. LeGeros, Calcium phosphates in oral biology and medicine. Karger, Basel, 1991. J.C. Elliot, Structure and chemistry of the apatites and other calcium orthophosphates, Vol. 18. Elsevier, Amsterdam, 1994. R.Z. LeGeros, Clin. Orthop. Rel. Res. 2002, 395, 81–98. D. Tadic, M. Epple, Biomaterials 2004, 25, 987–994. J.M. Rueger, W. Linhart, D. Sommerfeldt, Orthopa¨de 1998, 27, 89–95. Y. Doi, H. Iwanaga, T. Shibutani, Y. Morikawa, Y. Iwayama, J. Biomed. Mater. Res. 1999, 47, 424–433.

17 D. Briem, W. Linhart, W. Lehmann,

18

19

20

21

22

23 24

25 26

27

28

29

30 31

N.M. Meenen, J.M. Rueger, Unfallchirurg 2002, 105, 128–133. A.F. Schilling, W. Linhart, S. Filke, M. Gebauer, T. Schinke, J.M. Rueger, M. Amling, Biomaterials 2004, 25, 3963–3972. R. Weinkamer, M.A. Hartmann, Y. Brechet, P. Fratzl, Phys. Rev. Lett. 2004, 93, 228102. P. Ducy, M. Amling, S. Takeda, M. Priemel, A.F. Schilling, F.T. Beil, J. Shen, C. Vinson, J.M. Rueger, G. Karsenty, Cell 2000, 100, 197–207. A.F. Schilling, S. Filke, S. Brink, H. Korbmacher, M. Amling, J.M. Rueger, Eur. J. Trauma 2006, 32, 107–113. Y. Doi, T. Shibutani, Y. Moriwaki, T. Kajimoto, Y. Iwayama, J. Biomed. Mater. Res. 1998, 39, 603–610. G. Vaes, Clin. Orthoped. Rel. Res. 1988, 231, 239–271. S.L. Teitelbaum, M.M. Tondravi, F.P. Ross, J. Leucocyte Biol. 1997, 61, 381– 388. S.L. Teitelbaum, Science 2000, 289, 1504–1508. P. Koutsoukos, Z. Amjad, M.B. Tomson, G.H. Nancollas, J. Am. Chem. Soc. 1980, 102, 1553–1557. S. Yamada, D. Heymann, J.M. Bouler, G. Daculsi, Biomaterials 1997, 18, 1037–1041. R.P. Shellis, A.R. Lee, R.M. Wilson, J. Coll. Interface Sci. 1999, 218, 351– 358. A. Leeuwenburgh, P. Layrolle, F. Barre`re, J.D. de Bruijn, J. Schoonman, C.A. van Blitterswijk, K. de Groot, J. Biomed. Mater. Res. 2001, 56, 208–215. D. Tadic, M. Epple, Eur. J. Trauma 2002, 28, 136–137. C.R. Hankermeyer, K.L. Ohashi, D.C. Delaney, J. Ross, B.R. Constantz, Biomaterials 2002, 23, 743–750.

References 32 M.T. Fulmer, I.C. Ison, C.R.

33 34

35

36 37

38 39

40 41

42

43

44

45

46

47 48

49

Hankermayer, B.R. Constantz, J. Ross, Biomaterials 2002, 23, 751– 755. F. Peters, D. Reif, Mat.-wiss. u. Werkstofftech. 2004, 35, 203–207. S. Kamakura, Y. Sasano, T. Shimizu, K. Hatori, O. Suzuki, M. Kagayama, K. Motegi, J. Biomed. Mater. Res. 2002, 59, 29–34. J.C. Merry, I.R. Gibson, S.M. Best, W. Bonfield, J. Mater. Sci. Mater. Med. 1998, 9, 779–783. I.R. Gibson, W. Bonfield, J. Biomed. Mater. Res. 2002, 59, 697–708. W. Suchanek, P. Shuk, K. Byrappa, R.E. Riman, K.S. TenHuisen, V.F. Janas, Biomaterials 2002, 23, 699–710. D. Tadic, F. Peters, M. Epple, Biomaterials 2002, 23, 2553–2559. L.M. Rodriguez-Lorenzo, M. ValletRegi, J.M.F. Ferreira, Biomaterials 2001, 22, 583–588. D. Tadic, M. Epple, Biomaterials 2003, 24, 4565–4571. D. Tadic, A. Veresov, V.I. Putlayev, M. Epple, Mat.-wiss. u. Werkstofftech. 2003, 34, 1048–1051. D. Tadic, F. Beckmann, K. Schwarz, M. Epple, Biomaterials 2004, 25, 3335–3340. W. Linhart, D. Briem, M. Amling, J.M. Rueger, J. Windolf, Unfallchirurg 2004, 107, 154–157. B.R. Constantz, I.C. Ison, M.T. Fulmer, R.D. Poser, S.T. Smith, M. VanWagoner, J. Ross, S.A. Goldstein, J.B. Jupiter, D.I. Rosenthal, Science 1995, 267, 1796–1799. E. Fernandez, F.J. Gil, M.P. Ginebra, F.C.M. Driessens, J.A. Planell, S.M. Best, J. Mater. Sci. Mater. Med. 1999, 10, 169–176. E. Fernandez, F.J. Gil, M.P. Ginebra, F.C.M. Driessens, J.A. Planell, S.M. Best, J. Mater. Sci. Mater. Med. 1999, 10, 177–183. L.L. Hench, J. Am. Ceram. Soc. 1998, 81, 1705–1728. A.A. Ignatius, C. Schmidt, D. Kaspar, L.E. Claes, J. Biomed. Mater. Res. 2001, 55, 285–294. T. Welzel, W. Meyer-Zaika, M. Epple, Chem. Commun. 2004, 1204–1205.

50 A. Tampieri, G. Celotti, E. Landi,

51

52

53 54

55

56

57

58 59

60

61

62 63

64

65 66

Anal. Bioanal. Chem. 2005, 381, 568– 576. Y. Doi, T. Horiguchi, Y. Moriwaki, H. Kitago, T. Kajimoto, Y. Iwayama, J. Biomed. Mater. Res. 1996, 31, 43–49. C. Du, F.Z. Cui, Q.L. Feng, X.D. Zhu, K. de Groot, J. Biomed. Mater. Res. 1998, 42, 540–548. J.D. Hartgerink, E. Beniash, S.I. Stupp, Science 2001, 294, 1684–1688. D. Lickorish, J.A.M. Ramshaw, J.A. Werkmeister, V. Glattauer, C.R. Howlett, J. Biomed. Mater. Res. 2004, 68A, 19–27. S. Liao, F. Watari, M. Uo, S. Ohkawa, K. Tamura, W. Wang, F.Z. Cui, J. Biomed Mater. Res. Appl. Biomater. 2005, 74B, 817–821. H. Ehrlich, T. Douglas, D. Scharnweber, T. Hanke, R. Born, S. Bierbaum, H. Worch, Z. allg. anorg. Chem. 2005, 631, 1825–1830. A. Yokoyama, M. Gelinsky, T. Kawasaki, T. Kohgo, U. Koenig, W. Pompe, F. Watari, J. Biomed. Mater. Res. Appl. Biomater. 2005, 75B, 464– 472. E. Sachlos, D. Gotora, J.T. Czernuska, Tissue Eng. 2006, 12, 2479–2487. H. Peterlik, P. Roschger, K. Klaushofer, P. Fratzl, Nature Mater. 2006, 5, 52–55. H.S. Gupta, U. Stachewicz, W. Wagermaier, P. Roschger, H.D. Wagner, P. Fratzl, J. Mater. Res. 2006, 21, 1913–1921. M.A. Rauschmann, T.A. Wichelhaus, V. Stirnal, E. Dingeldein, L. Zichner, R. Schnettler, V. Alt, Biomaterials 2005, 26, 2677–2684. P. Westbrook, F. Marin, Nature 1998, 392, 861–862. Y.W. Kim, J.J. Kim, Y.H. Kim, J.Y. Rho, Biomaterials 2001, 23, 2089– 2096. A.G. Mikos, A.J. Thorsen, L.A. Czerwonka, Y. Bao, R. Langer, D.N. Winslow, J.P. Vacanti, Polymer 1994, 35, 1068–1077. R. Langer, Acc. Chem. Res. 2000, 33, 94–101. R. Langer, N.A. Peppas, A. I. Ch. E. J. 2003, 49, 2990–3006.

93

94

6 Biomimetic Bone Substitution Materials ¨ stman, H. Pihlajama¨ki, 67 O. Bo 68

69 70

71

72 73 74 75 76

77 78

79

80

81

82 83

84

Biomaterials 2000, 21, 2615–2621. F. von Burkersroda, L. Schedl, A. Go¨pferich, Biomaterials 2002, 23, 4221–4231. K.D. Ku¨hn, Bone cements. Springer, Berlin, 2000. B.D. Ratner, A.S. Hoffman, F.J. Schoen, Biomaterials Science. An Introduction to Materials in Medicine. Academic Press, 2004. M. Vert, G. Schwarch, J. Coudane, J. Macromol. Sci. Pure Appl. Chem. 1995, A32, 787–796. N.A. Ashammakhi, J. Biomed. Mater. Res. 1996, 33, 297–303. O.M. Bo¨stman, Clin. Orthop. Rel. Res. 1996, 329, 233–239. E. Chiellini, R. Solaro, Adv. Mater. 1996, 8, 305–313. N. Ashammakhi, P. Rokkanen, Biomaterials 1997, 18, 3–9. S.L. Ishaug-Riley, G.M. Crane, A. Gurlek, M.J. Miller, A.W. Yasko, M.J. Yaszemski, A.G. Mikos, J. Biomed. Mater. Res. 1997, 36, 1–8. K. Schwarz, M. Epple, Macromol. Rapid Commun. 1998, 19, 613–617. L. Lu, S.J. Peter, M.D. Lyman, H.L. Lai, S.M. Leite, J.A. Tamada, J.P. Vacanti, R. Langer, A.G. Mikos, Biomaterials 2000, 21, 1595–1605. C. Schiller, C. Rasche, M. Wehmo¨ller, F. Beckmann, H. Eufinger, M. Epple, S. Weihe, Biomaterials 2004, 25, 1239–1247. H. Eufinger, C. Rasche, J. Lehmbrock, M. Wehmo¨ller, S. Weihe, I. Schmitz, C. Schiller, M. Epple, Biomaterials 2007, 28, 475–485. C.K. Breuer, T. Shin’oka, R.E. Tanel, G. Zund, D.J. Mooney, P.X. Ma, T. Miura, S. Colan, R. Langer, J.E. Mayer, J.P. Vacanti, Biotechn. Bioeng. 1996, 50, 562–567. K.C. Dee, R. Bizios, Biotechn. Bioeng. 1996, 50, 438–442. L.E. Niklason, J. Gao, W.M. Abbott, K.K. Hirschi, S. Houser, R. Marini, R. Langer, Science 1999, 284, 489– 493. Q.Q. Qiu, P. Ducheyne, P.S. Ayyaswamy, J. Biomed. Mater. Res. 2000, 52, 66–76.

85 R. Zhang, P.X. Ma, J. Biomed. Mater.

Res. 2000, 52, 430–438. 86 C. Martin, H. Winet, J.Y. Bao,

Biomaterials 1996, 17, 2373–2380. 87 H. Winet, J.Y. Bao, J. Biomater. Sci.

Polymer Edn. 1997, 8, 517–532. 88 C.M. Agrawal, K.A. Athanasiou,

89

90 91

92 93

94

95

96

97

98

99

100

J. Biomed. Mater. Res. 1997, 38, 105– 114. W. Linhart, F. Peters, W. Lehmann, A.F. Schilling, K. Schwarz, M. Amling, J.M. Rueger, M. Epple, J. Biomed. Mater. Res. 2001, 54, 162– 171. C. Schiller, M. Epple, Biomaterials 2003, 24, 2037–2043. W. Linhart, W. Lehmann, M. Siedler, F. Peters, A.F. Schilling, K. Schwarz, M. Amling, J.M. Rueger, M. Epple, J. Mater. Sci. 2006, 41, 4806–4813. G. Ryan, A. Pandit, D.P. Apatsidis, Biomaterials 2006, 27, 2651–2670. M. Bram, A. Laptev, H.P. Buchkremer, D. Sto¨ver, Mat.-wiss. u. Werkstofftech. 2004, 35, 213–218. M. Assad, P. Jarzem, M.A. Leroux, C. Coillard, A.V. Chernyshov, S. Charette, C.H. Rivard, J. Biomed. Mater. Res. Appl. Biomater. 2003, 64B, 107–120. M. Assad, A.V. Chernyshov, P. Jarzem, M.A. Leroux, C. Coillard, S. Charette, C.H. Rivard, J. Biomed. Mater. Res. Appl. Biomater. 2003, 64B, 121–129. M. Assad, F. Likibi, P. Jarzem, M.A. Leroux, C. Coillard, C.H. Rivard, Mat.-wiss. u. Werkstofftech. 2004, 35, 219–223. O. Prymak, D. Bogdanski, M. Ko¨ller, S.A. Esenwein, G. Muhr, F. Beckmann, T. Donath, M. Assad, M. Epple, Biomaterials 2005, 26, 5801– 5807. H. Matsuno, A. Yokoyama, F. Watari, M. Uo, T. Kawasaki, Biomaterials 2001, 22, 1253–1262. F. Witte, J. Reifenrath, P.P. Mu¨ller, H.A. Crostack, J. Nellesen, F.W. Bach, D. Bormann, M. Rudert, Mat.-wiss. u. Werkstofftech. 2006, 37, 504–508. T. Kokubo, Acta Mater. 1998, 46, 2519–2527.

References 101 P. Habibovic, F. Barre`re, C.A. van

102

103

104

105

106

107 108 109

110 111 112

113

114

115

116

Blitterswijk, K. de Groot, P. Layrolle, J. Am. Ceram. Soc. 2002, 85, 517– 522. Y. Han, K. Xu, G. Montay, T. Fu, J. Lu, J. Biomed. Mater. Res. 2002, 60, 511–516. L. Sun, C.C. Berndt, K.A. Khor, H.N. Cheang, K.A. Gross, J. Biomed. Mater. Res. 2002, 62, 228–236. J. Choi, D. Bogdanski, M. Ko¨ller, S.A. Esenwein, D. Mu¨ller, G. Muhr, M. Epple, Biomaterials 2003, 24, 3689– 3696. H. Schliephake, D. Scharnweber, M. Dard, S. Ro¨ßler, A. Sewing, C. Hu¨ttmann, J. Biomed. Mater. Res. 2003, 64A, 225–234. J.W.M. Vehof, J. van den Dolder, J.E. de Ruijter, P.H.M. Spauwen, J.A. Jansen, J. Biomed. Mater. Res. 2003, 64A, 417–426. S. Weiner, L. Addadi, H.D. Wagner, Mater. Sci. Eng. 2000, C11, 1–8. M. Neumann, M. Epple, Eur. J. Trauma 2006, 32, 125–131. H. Liao, H. Mutvei, L. Hammarstro¨m, T. Wurtz, J. Li, Biomaterials 2002, 23, 2693–2701. M. Ni, B.D. Ratner, Biomaterials 2003, 24, 4323–4331. U. Ripamonti, S. Ma, A.H. Reddi, Matrix 1992, 12, 202–212. J. Vuola, T. Bo¨hling, H. Go¨ransson, P. Puolakkainen, J. Biomed. Mater. Res. 2002, 59, 152–159. C.N. Demers, M. Tabrizian, A. Petit, R.C. Hamdy, H. L’Yahia, J. Biomed. Mater. Res. 2002, 59, 403–410. Y. Acil, I.N.G. Springer, V. Broek, H. Terheyden, S. Jepsen, J. Cell. Biochem. 2002, 86, 90–98. D. Turhani, E. Watzinger, M. Weißenbo¨ck, B. Cvikl, D. Thurnher, G. Wittwer, K. Yerit, R. Ewers, J. Oral Maxillofac. Surg. 2005, 63, 673–681. R. Ewers, J. Oral Maxillofac. Surg. 2005, 63, 1712–1723.

117 F. Wanschitz, M. Figl, A. Wagner, R.

118 119

120

121 122 123

124

125

126

127

128 129

130 131

Ewers, Int. J. Oral Maxillofac. Surg. 2006, 21, 433–438. K.A. Gross, E. Ezerietis, J. Biomed. Mater. Res. 2003, 64A, 672–683. A. Tampieri, G. Celotti, S. Sprio, A. Delcogliano, S. Franzese, Biomaterials 2001, 22, 1365–1370. W. Pompe, H. Worch, M. Epple, W. Friess, M. Gelinsky, P. Greil, U. Hempel, D. Scharnweber, K. Schulte, Mater. Sci. Eng. 2003, A362, 40–60. A.H. Reddi, Nature Biotechnol. 1998, 16, 247–252. E.H.J. Groeneveld, E.H. Burger, Eur. J. Endocrinol. 2000, 142, 9–21. P.H. Warnke, I.N. Springer, J. Wiltfang, Y. Acil, H. Eufinger, M. Wehmo¨ller, P.A.J. Russo, H. Bolte, E. Sherry, E. Behrens, H. Terheyden, Lancet 2004, 364, 766–770. M. Kantlehner, D. Finsinger, J. Meyer, P. Schaffner, A. Jonczyk, B. Diefenbach, B. Nies, H. Kessler, Angew. Chem. 1999, 111, 587–590. J.A. Rowley, Z. Sun, D. Goldman, D.J. Mooney, Adv. Mater. 2002, 14, 886–889. J. Gille, B. Dorn, J. Kekow, J. Bruns, P. Behrens, Int. Orthopaed. 2002, 26, 203–206. L. Meseguer-Olmo, M.J. Ros-Nicola´s, M. Clavel-Sainz, V. Vicente-Ortega, M. Alcaraz-Banos, A. Lax-Pe´rez, D. Arcos, C.V. Ragel, M. Vallet-Regi, J. Biomed. Mater. Res. 2002, 61, 458– 465. H.W. Kim, J.C. Knowles, H.E. Kim, Biomaterials 2004, 25, 1279–1287. D. Tadic, T. Welzel, P. Seidel, E. Wu¨st, E. Dingeldein, M. Epple, Mat.wiss. u. Werkstofftech. 2004, 35, 1001– 1005. M. Vallet-Regi, Dalton Trans. 2006, 5211–5220. M. Wehmo¨ller, K. Neuking, M. Epple, T. Annen, H. Eufinger, Mat.-wiss. u. Werkstofftech. 2006, 413–415.

95

97

7 Simulated Body Fluid (SBF) as a Standard Tool to Test the Bioactivity of Implants Tadashi Kokubo and Hiroaki Takadama

Abstract

Most bone-bonding bioactive materials form bone-like apatite on their surfaces after being implanted into the living body, and bond to neighboring bone through this apatite layer. The apatite layer can be reproduced on the surfaces of materials in an organic-substance-free acellular simulated body fluid (SBF) with ion concentrations almost equal to those of human blood plasma. The bone-bonding ability of a material is often evaluated by examining the ability of apatite to form on the material in SBF. In this chapter, the validity of this method for evaluating the bone-bonding bioactivity of a material, the ion concentrations of SBF, the materials able to form apatite, the characteristics of apatite, the bone-bonding mechanisms of bioactive materials, and the mechanisms of apatite formation, are reviewed. Key words: simulated body fluid (SBF), bioactive material, apatite-forming ability, bone-bonding ability, bone substitute, bone-like apatite.

7.1 Introduction

Various materials, including Bioglass [1], sintered hydroxyapatite [2], sintered beta-tricalcium phosphate [3], biphasic ceramics of hydroxyapatite and tricalcium phosphate [4], and glass–ceramic A-W [A ¼ apatite (Ca10 (PO4 )6 (O, F2 )); W ¼ wollastonite (CaOSiO2 )] [5], can bond to living bone. These are referred to as ‘‘bioactive’’ materials, and many are currently in clinical use as important bone substitutes. Most of them bond to living bone through an apatite layer that forms on their surfaces after implantation into the living body. This apatite formation has been reproduced on their surfaces in an organic-substance-free acellular simulated body fluid (SBF), with ion concentrations almost equal to those of human blood plasma [6]. This indicates that the bone-bonding bioactivity of a Handbook of Biomineralization. Edited by M. Epple and E. Baüerlein Copyright 8 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ISBN: 978-3-527-31806-3

98

7 Simulated Body Fluid (SBF) as a Standard Tool to Test the Bioactivity of Implants

material can be evaluated in preliminary fashion, before conducting animal experiments, by examining apatite formation on its surface in SBF. As a result, the number of animals required to evaluate the bone-bonding ability of a material can be reduced, and today many laboratories utilize SBF as a standard tool when testing the bioactivity of new materials. In this chapter, the correlation between the bone-bonding bioactivity of materials and their apatite-forming ability in SBF, the ion concentrations of SBF, the dependence of apatite formation on the material, the characteristics of apatite, the bone-bonding mechanisms of bioactive materials, and the mechanisms of apatite formation on these materials, are described.

7.2 Qualitative Correlation of Bone-Bonding Bioactivity of a Material with Apatite Formation on its Surface in SBF

Materials that bond to living bone through an apatite or calcium phosphate layer that forms on their surfaces after implantation into the living body include: a Bioglass 45S5 type-glass in a Na2 OaCaOaSiO2 aP2 O5 system [7]; bioactive glasses in the Na2 OaCaOaB2 O3 aAl2 O3 aP2 O5 system [8]; glasses in the CaOaSiO2 system [9]; Ceravital-type glass–ceramics containing crystalline apatite in the Na2 OaCaOaSiO2 aP2 O5 system [10]; Glass–ceramic A-W, containing crystalline apatite and wollastonite, in the MgOaCaOaSiO2 aP2 O5 system [11]; Bioverite-type glass–ceramics containing crystalline apatite and fluorophlogopite in the Na2 OaMgOaCaOaAl2 O3 aSiO2 aP2 O5 aF system [12]; sintered hydroxyapatite [13]; biphasic ceramics of hydroxyapatite and beta-tricalcium phosphate; sintered calcium sulfate [14]; a composite of glass–ceramic A-W with polyethylene [15]; titanium metal subjected to NaOH and heat treatments [16]; and tantalum metal subjected to NaOH and heat treatments [17]. An example of an interface of glass–ceramic A-W to living bone is shown in Figure 7.1. All of these bone-bonding bioactive glasses, glass–ceramics, sintered crystalline ceramics, composites and metals have been confirmed as forming an apatite on their surfaces in SBF within 4 weeks [4, 6–8, 10, 14, 18–23], except for the Bioverite-type glass–ceramic, which has not been investigated for apatite forma-

7.2 Qualitative Correlation of Bone-Bonding Bioactivity of a Material with Apatite Formation

Fig. 7.1 Transmission electron microscopy image of the interface of glass–ceramic A-W and rabbit tibial bone [13].

tion on its surface in SBF. An apatite layer formed on a glass–ceramic A-W in SBF is illustrated in Figure 7.2. When a small amount of Al2 O3 was added to the composition of Bioglass-type glass [24], CaOaSiO2 glass [25], and glass–ceramic A-W [26], the resultant glasses and glass–ceramics did not form an apatite or calcium phosphate layer on their surfaces in the living body, and did not bond to the neighboring bone. In addition, none of these materials with added Al2 O3 formed apatite on their surfaces within 4 weeks in SBF [14, 27, 28]. It can be concluded from these results that the essential requirement for a material to bond to living bone is the formation of an apatite or calcium phosphate layer on its surface, and that the bone-bonding bioactivity of a material can be evaluated by examining the formation of an apatite layer on its surface in SBF. However, it should be noted here that a small number of cases in which a material bonds to living bone without yielding a detectable apatite layer at their interfaces have been reported. Sintered beta-tricalcium phosphate and a natural calcite of calcium carbonate are examples [29, 30], with neither material forming an apatite layer on its surface within 4 weeks in SBF [31, 32]. In fact, the bone-bonding properties of these materials might be related to their high resorbability in the living body. One case in which a material – abalone shell – does not bond to living bone, despite forming an apatite or calcium phosphate layer on its surface in the living

99

100


Fig. 7.2 The apatite layer formed on glass–ceramic A-W in simulated body fluid [19].

body – has been also reported. Abalone shell also forms an apatite layer on its surface in SBF [32], and suppression of the bone-bonding bioactivity of this material might be attributed to foreign body reactions elicited by small amounts of protein contained in the shell. From the above findings, it can be concluded that a material which is able to form an apatite layer on its surface in SBF may bond to living bone through the apatite layer that forms on its surface, as long as the material does not release any component that induces toxic or immune responses in the surrounding tissue. Based on these findings, the examination of apatite formation on a surface of a material in SBF would be a useful tool for predicting the bone-bonding bioactivity of a material, before progressing to animal experiments. Indeed, by using this method not only the number of animals but also the duration of animal experiments required to evaluate the bone-bonding bioactivity of a material can be greatly reduced.

7.3 Quantitative Correlation of Bone-Bonding Bioactivity and Apatite-Forming Ability in SBF

Not all bioactive materials show equal bone-bonding ability; rather, the time required for a material to bond to living bone, and the amount of bone formed

7.4 Ion Concentrations of SBF

Fig. 7.3 The rate of bone formation on a cross-section of a defect of rabbit femur when filled with glass particles 6 weeks after implantation compared with time of surface apatite formation in simulated body fluid [33].

around a material in a given time, will vary widely depending on the material involved. The time required for a bioactive material to become fully covered with apatite in SBF also varies, depending on the material. In order to investigate the relationship between bone formation in vivo and apatite formation in SBF, bone formation in defects in rabbit femurs filled with Na2 OaCaOaSiO2 glass particles (the SiO2 contents of which were changed from 70.0 to 50.0 mol%, with a constant Na2 O/CaO molar ratio of one) were examined [33]. The time required for the same glasses in SBF to form bone-like apatite which fully covered their surfaces was also measured [34]. The data provided in Figure 7.3 show clearly that bone formation around glass particles increases with the increasing apatiteforming ability of the glasses in SBF. This, in turn, indicates that the bonebonding bioactivity of a material can be evaluated not only qualitatively but also quantitatively, by examining the apatite-forming ability on the material’s surface in SBF.

7.4 Ion Concentrations of SBF

In all of the above-described investigations, the organic-substance-free acellular solution used as the SBF had ion concentrations as first reported by Kokubo et al. in 1990 [6], and as corrected by the same authors in 1991 [35]. However, the ion concentrations of this SBF were not exactly equal to those of human blood plasma (see Table 7.1), as SBF is richer in Cl ions and poorer in HCO3 ions than is human blood plasma [36]. In 2003, Oyane et al. proposed a revised simulated body fluid (r-SBF), in which the ion concentrations were identical to those of human blood plasma [37]. However, r-SBF had a strong tendency to produce

101

102

7 Simulated Body Fluid (SBF) as a Standard Tool to Test the Bioactivity of Implants Table 7.1 Ion concentrations of simulated body fluid (SBF) and human blood plasma.

Ion concentration [mmol]

Blood plasma SBF

NaB

KB

Mg 2B

Ca 2B

ClC

HCO3C

HPO4 2C

SO4 2C

142.0 142.0

5.0 5.0

1.5 1.5

2.5 2.5

103.0 147.8

27.0 4.2

1.0 1.0

0.5 0.5

precipitates of calcium carbonate, as it is highly supersaturated with respect to hydroxyapatite and calcite [38]. In 2004, the method for preparing conventional SBF was further refined and simplified such that it could be easily prepared and subjected to round-robin testing by 10 research institutes [39]. This refined SBF recipe (the details of which have been published [40]) was proposed to the International Organization for Standardization as a standard solution for in-vitro monitoring of the apatiteforming ability of implant materials. Simulated body fluids with higher ion concentrations (e.g., 1.5 and 4 SBF, where ion concentrations are 1.5- or four-times those of SBF) have also been used to evaluate the bone-bonding abilities of materials, or the production of a bone-like apatite layer on materials. It should be noted, however, that no correlation has been identified between apatite formation in such solutions and bone-bonding ability, and that the apatite formed in these solutions differs in composition from bone mineral [41].

7.5 Materials Able to Form Apatite

Despite human body fluid being highly supersaturated with respect to apatite (even under normal conditions [42]), apatite does not usually precipitate in the living body, except at sites of bone tissue, as the energy barrier for its nucleation is high. This means that, once apatite nuclei have been formed catalytically on a material, they can grow spontaneously by consuming the calcium and phosphate ions from the surrounding body fluid. The question persists, however, as to what type of material induces apatite nucleation. In an attempt to answer this question, various types of gels prepared using sol-gel methods were soaked in SBF, and their apatite-forming abilities examined. Although SiO2 [43], TiO2 [44], ZrO2 [45], Nb2 O5 [46] and Ta2 O5 [47] gels were seen to form apatite on their surfaces, Al2 O3 [44] gels did not (Fig. 7.4), which indicated that the SiaOH, TiaOH, ZraOH, NbaOH and TaaOH groups that were abundant on the surfaces of the gels were effective in inducing apatite

7.6 Composition and Structure of Apatite

Fig. 7.4 Apatite formed on (left) SiO2 and (right) TiO2 gels in simulated body fluid [44].

nucleation. Subsequently, Tanahashi et al., using self-assembled monolayers, showed that COOH and PO4 H2 groups were also effective for apatite nucleation [48]. Based on these findings, titanium metal and its alloys were subjected to NaOH solution and heat treatment to form sodium titanate on their surfaces. These treated materials were found to form bone-like apatite on their surfaces in the living body, and to bond to living bone [49], and subsequently were applied for use in hip-joint prostheses.

7.6 Composition and Structure of Apatite

The calcium phosphate layer formed on bioactive materials after implantation into living bodies has been identified by micro X-ray diffraction [50] and electron [51] diffraction as a nanosized crystalline apatite. However, it has been difficult to obtain more detailed structural information for these calcium phosphate layers formed in vivo. More detailed structural information can be obtained for apatite formed on bioactive materials in SBF. According to observations made with transmission electron microscopy (TEM), the apatite on both glass–ceramic A-W [52] and NaOH- and heat-treated titanium metal [53] in SBF takes the shape of thin needles of 10 nm thickness and 100 nm length (Fig. 7.5). This apatite has a Ca/P atomic ratio of about 1.65, which is less than the stoichiometric value of 1.67, and contains a small amount of Naþ and Mg 2þ ions beside CO3 2 ions [19, 52, 53]. As these characteristics are similar to those of bone mineral, the material may be referred to as ‘‘bone-like’’ apatite.

103

104


Fig. 7.5 Transmission electron microscopy (left) and energy dispersive X-ray analysis (right) images of apatite formed on NaOH- and heattreated titanium metal in simulated body fluid [53].

7.7 Mechanism of Bonding of Bioactive Material to Bone

As described above, most bioactive materials form apatite on their surfaces after being implanted into the living body. As this surface apatite is very similar to bone mineral in its composition, structure and morphology, the bone-producing cells (osteoblasts) could preferentially proliferate and differentiate on its surface to produce collagen and apatite, similar to their behavior on the surface of fractured bone (Fig. 7.6) [54]. As a result, the surrounded bone may come into direct

Fig. 7.6 Transmission electron microscopy image of the interface of glass–ceramic A-W and rabbit tibial bone at early stage after implantation [54].

7.8 Mechanisms of Apatite Formation

contact with the surface apatite layer on materials. When this occurs, a tight chemical bond is formed between the apatite in the bone and the surface apatite to reduce their interface energy.

7.8 Mechanisms of Apatite Formation

If apatite formation on bioactive materials implanted into the living body can be reproduced on their surfaces in SBF, then the mechanisms of apatite formation on the materials might be revealed by the examining surface structural changes of the materials as a function of soaking time in SBF. Based on TEM observations and zeta potential measurements, the mechanism of apatite formation on sintered hydroxyapatite in body environment is interpreted as follows [55]. The sintered hydroxyapatite is initially negatively charged on its surface, and combines with positively charged Ca 2þ ions in the surrounding fluid. As a result, Ca-rich amorphous calcium phosphate is formed on the sintered hydroxyapatite. As the Ca 2þ ions accumulate, the sintered hydroxyapatite becomes positively charged on its surface and reacts with negatively charged phosphate ions. As a result, Ca-poor amorphous calcium phosphate is formed which is eventually transformed into the more stable, nanosized crystalline bone-like apatite. This mechanism is essentially the same in fluids containing proteins [56]. Apatite formation on NaOH- and heat-treated titanium metal in a body environment is similarly interpreted as follows [53, 57]. The treated titanium metal releases Naþ ions from its surface sodium titanate layer via exchange with H3 Oþ ions in the fluid, to form TiaOH groups (Fig. 7.7). As a result, its surface becomes negatively charged and reacts with positively charged Ca 2þ ions to form calcium titanate. As the calcium ions accumulate, the positively charged surface reacts with negatively charged phosphate ions, forming amorphous calcium phosphate. As this phase is metastable, it eventually transforms into nanosized, crystalline bone-like apatite.

Fig. 7.7 The mechanism of bone-like apatite formation on NaOH- and heat-treated titanium metal in vitro [57]. SBF ¼ simulated body fluid.

105

106


7.9 Summary

Simulated body fluid, which is used to test the bone-bonding bioactivity of various materials, is identical to human blood plasma in terms of its ion concentrations, but does not contain organic substances such as proteins. Nevertheless, the soaking of bioactive materials in SBF can reproduce the apatite formation seen on such materials in the living body. SBF is easily prepared and relatively stable at body temperature. Moreover, it is useful for evaluating the bone-bonding bioactivity of new materials and investigating the mechanisms of apatite formation on, and bone bonding of, materials.

References 1 L.L. Hench, R.J. Splinter, W.C. Allen,

2

3 4

5

6

7

8 9

10

11

T.K. Greenlee, J. Biomed. Mater. Res. 1972, 2, 117–141. M. Jarcho, J.I. Kay, R.H. Gummaer, H.P. Drobeck, J. Bioeng. 1977, 1, 79–92. B.J. Rejda, J.G.J. Peelen, K. de Groot, J. Bioeng. 1977, 1, 93–97. R.Z. Legeros, S. Lin, R. Rohanizadeh, D. Mijares, J.P. Legeros, J. Mater. Sci. Mater. Med. 2003, 14, 201–209. T. Kokubo, M. Shigematsu, Y. Nagashima, M. Tashiro, T. Nakamura, T. Yamamuro, S. Higashi, Bull. Inst. Chem. Res. Kyoto Univ. 1982, 60, 260–268. T. Kokubo, H. Kushitani, S. Sakka, T. Kitsugi, T. Yamamuro, J. Biomed. Mater. Res. 1990, 24, 721–734. M. Ogino, F. Ohuchi, L.L. Hench, J. Biomed. Mater. Res. 1980, 14, 55–64. ¨ .H. Anderson, K.H. Karlsson, J. NonO Cryst. Solids 1991, 129, 145–151. K. Ohura, T. Nakamura, T. Yamamuro, T. Kokubo, Y. Ebisawa, Y. Kotoura, M. Oka, J. Biomed. Mater. Res. 1991, 25, 357–365. C. Ohtsuki, H. Kushitani, T. Kokubo, S. Kotani, T. Yamamuro, J. Biomed. Mater. Res. 1991, 25, 1363–1370. T. Kitsugi, T. Nakamura, T. Yamamuro, T. Kokubo, T. Shibuya, M. Takagi, J. Biomed. Mater. Res. 1987, 21, 1255–1271.

¨ land, W. Vogel, K. Naumann, 12 W. Ho

13

14

15

16

17

18 19

J. Biomed. Mater. Res. 1985, 19, 303–312. M. Neo, S. Kotani, T. Nakamura, T. Yamamuro, C. Ohtsuki, T. Kokubo, Y. Bando, J. Biomed. Mater. Res. 1992, 26, 1419–1432. H. Chan, D. Mijares, J.L. Ricci, Transactions of the Seventh World Biomaterials Congress. Australian Society for Biomaterials Inc., Brunswick Lower, Victoria, Australia, 2004, p. 627. J.A. Juhasz, S. Ishii, S.M. Best, M. Kawashita, M. Neo, T. Kokubo, T. Nakamura, W. Bonfield, Transactions of the Seventh World Biomaterials Congress, Australian Society for Biomaterials Inc., Brunswick Lower, Victoria, Australia, 2004, p. 665. S. Nishiguchi, S. Fujibayashi, H.-M. Kim, T. Kokubo, T. Nakamura, J. Biomed. Mater. Res. 2003, 67A, 28–35. H. Kato, T. Nakamura, S. Nishiguchi, Y. Matsusue, M. Kobayashi, T. Miyazaki, H.-M. Kim, T. Kokubo, J. Biomed. Mater. Res. Appl. Biomater. 2000, 53, 28–35. C. Ohtsuki, T. Kokubo, T. Yamamuro, J. Non-Cryst. Solids 1992, 143, 84–92. T. Kokubo, S. Ito, T. Huang, T. Hayashi, S. Sakka, T. Kitsugi, T. Yamamuro, J. Biomed. Mater. Res. 1990, 24, 331–343.

References 20 T. Kokubo, M. Kushiyani, Y. Ebisawa,

21

22

23

24

25

26

27

28

29

30

31

32

33

34

T. Kitsugi, S. Kotani, K. Oura, T. Yamamuro, Bioceramics, Volume 1. Ishiyaku EuroAmerica, Tokyo, 1988, pp. 157–162. J.A. Judasz, S.M. Best, W. Bonfield, M. Kawashita, N. Miyata, T. Kokubo, T. Nakamura, J. Mater. Sci. Mater. Med. 2003, 14, 489–495. H.-M. Kim, F. Miyaji, T. Kokubo, T. Nakamura, J. Biomed. Mater. Res. 1996, 32, 409–417. T. Miyazaki, H.-M. Kim, F. Miyaji, T. Kokubo, T. Nakamura, J. Biomed. Mater. Res. 2000, 50, 35–42. ¨ .H. Andersson, G. Liu, K.H. O Karlsson, L. Niemi, J. Miettinen, J. Juhanoja, J. Mater. Sci. Mater. Med. 1990, 1, 219–227. K. Ohura, T. Nakamura, T. Yamamuro, Y. Ebisawa, T. Kokubo, Y. Kotoura, M. Oka, J. Mater. Sci. Mater. Med. 1992, 3, 95–100. T. Kitsugi, T. Yamamuro, T. Nakamura, T. Kokubo, Int. Orthop. 1989, 13, 199–206. ¨ .H. Andersson, G. Liu, K. O Kangasniemi, J. Juhanoja, J. Mater. Sci. Mater. Med. 1992, 3, 145–150. Y. Ebisawa, T. Kokubo, K. Ohura, T. Yamamuro, J. Mater. Sci. Mater. Med. 1990, 1, 239–244. S. Kotani, Y. Fujita, T. Kitsugi, T. Nakamura, T. Yamamuro, J. Biomed. Mater. Res. 1991, 25, 1303–1315. Y. Fujita, T. Yamamuro, T. Nakamura, S. Kotani, J. Biomed. Mater. Res. 1991, 25, 991–1003. C. Ohtsuki, T. Kokubo, M. Neo, S. Kotani, T. Yamamuro, T. Nakamura, Y. Bando, Phos. Res. Bull. 1991, 1, 191–196. C. Ohtsuki, Y. Aoki, T. Kokubo, Y. Fujita, S. Kotani, T. Yamamuro, Transactions of the 11th Annual Meeting of Japanese Society for Biomaterials. Japanese Society for Biomaterials, Toshima-ku, Tokyo, Japan, 1989, p. 12. S. Fujibayashi, M. Neo, H.-M. Kim, T. Kokubo, T. Nakamura, Biomaterials 2003, 24, 1349–1356. H.-M. Kim, F. Miyaji, T. Kokubo, C. Ohtsuki, T. Nakamura, J. Am. Ceram. Soc. 1995, 78, 2405–2411.

35 T. Kokubo, Biomaterials 1991, 12,

155–163. 36 J.E. Gamble, Chemical Anatomy,

37

38

39

40 41

42

43

44

45

46

47

48 49

50

51

Physiology and Pathology of Extracellular Fluid. Harvard University Press, Cambridge, 1967, pp. 1–17. A. Oyane, H.-M. Kim, T. Furuya, T. Kokubo, T. Miyazaki, T. Nakamura, J. Biomed. Mater. Res. 2003, 65A, 188–195. A. Oyane, K. Onuma, A. Ito, H.-M. Kim, T. Kokubo, T. Nakamura, J. Biomed. Mater. Res. 2003, 64A, 339–348. H. Takadama, M. Hashimoto, M. Mizuno, T. Kokubo, Phos. Res. Bull. 2004, 17, 119–125. T. Kokubo, H. Takadama, Biomaterials 2006, 27, 2907–2915. H.-M. Kim, K. Kishimoto, F. Miyaji, T. Kokubo, T. Yao, Y. Suetsugu, J. Tanaka, T. Nakamura, J. Biomed. Mater. Res. 1999, 46, 228–235. W. Neuman, M. Neuman, The Chemical Dynamics of Bone Mineral. University of Chicago, 1958, p. 34. P. Li, C. Ohtsuki, T. Kokubo, K. Nakanishi, N. Soga, T. Nakamura, J. Am. Ceram. Soc. 1992, 75, 2094– 2097. P. Li, C. Ohtsuki, T. Kokubo, K. Nakanishi, N. Soga, K. de Groot, J. Biomed. Mater. Res. 1994, 28, 7–15. M. Uchida, H.-M. Kim, T. Kokubo, T. Nakamura, J. Am. Ceram. Soc. 2001, 84, 2041–2044. T. Miyazaki, H.-M. Kim, T. Kokubo, C. Ohtsuki, H. Kato, T. Nakamura, J. Ceram. Soc. Japan 2001, 109, 929–933. T. Miyazaki, H.-M. Kim, T. Kokubo, H. Kato, T. Nakamura, J. Sol-Gel Sci. Technol. 2001, 21, 83–88. M. Tanahashi, T. Matsuda, J. Biomed. Mater. Res. 1997, 34, 305–315. T. Kokubo, H.-M. Kim, M. Kawashita, T. Nakamura, J. Mater. Sci. Mater. Med. 2004, 15, 99–107. T. Kokubo, C. Ohtsuki, S. Kotani, T. Kitsugi, T. Yamamuro, Bioceramics, Volume 2. German Ceramic Society, Cologne, 1990, pp. 113–121. M. Neo, S. Kotani, Y. Fujita, T. Nakamura, T. Yamamuro, Y. Bando,

107

108

7 Simulated Body Fluid (SBF) as a Standard Tool to Test the Bioactivity of Implants C. Ohtsuki, T. Kokubo, J. Biomed. Mater. Res. 1992, 26, 255–267. 52 C. Ohtsuki, Y. Aoki, T. Kokubo, Y. Bando, M. Neo, T. Nakamura, J. Ceram. Soc. Japan 1995, 103, 449–454. 53 H. Takadama, H.-M. Kim, T. Kokubo, T. Nakamura, J. Biomed. Mater. Res. 2001, 57, 441–448. 54 M. Neo, T. Nakamura, C. Ohtsuki, T. Kokubo, T. Yamamuro, J. Biomed. Mater. Res. 1993, 27, 999–1006.

55 H.-M. Kim, T. Himeno, M.

Kawashita, T. Kokubo, T. Nakamura, J. R. Soc. Interface 2004, 1, 17–22. 56 T. Kokubo, T. Himeno, H.-M. Kim, M. Kawashita, T. Nakamura, Bioceramics, Volume 16, Trans. Tech. Pub., Switzerland, 2004, pp. 139–142. 57 H.-M. Kim, T. Himeno, M. Kawashita, J.H. Lee, T. Kokubo, T. Nakamura, J. Biomed. Mater. Res. 2003, 67A, 1305–1309.

109

8 Stimulation of Bone Growth on Implants by Integrin Ligands Mońica Lo´pez-Garcıá and Horst Kessler

Abstract

A successful biointegration of orthopedic and craniofacial implants requires a strong mechanical interaction between the surface of the artificial material and the surrounding natural bone tissue. Osseointegration of implants is known to be a biological process that occurs by formation of new peri-implant bone in direct contact with the synthetic surface. Mimicking the physiological adhesion process of osteoblasts to the extracellular matrix (ECM), by coating of implant surfaces with specific cell-adhesive molecules, was proven to enhance osteoblast adhesion in vitro and to accelerate osseointegration of implants in vivo. Cell adhesion is mediated by integrins, a class of heterodimeric transmembrane cell receptors that bind selectively to different proteins of the ECM and transduce information to the nucleus through cytoplasmic signaling pathways. The peptide sequence Arg-Gly-Asp (RGD), is by far the most effective and extensively studied ligand to promote osteoblast adhesion and proliferation on implants through integrin stimulation. The biofunctionalization of different surfaces with RGD peptides and mimetics has resulted in major improvements in bone implant technology. Key words: integrins, RGD peptides, cell adhesion, surface coating, implants, osseointegration.

8.1 Introduction 8.1.1 Biomimetic Materials for Implant Technology

Biomaterials are designed to restore or replace a damaged part of the body and/or its associated functions. They may be used either in a permanent way or as temHandbook of Biomineralization. Edited by M. Epple and E. Baüerlein Copyright 8 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ISBN: 978-3-527-31806-3

110

8 Stimulation of Bone Growth on Implants by Integrin Ligands

porary support for cells and tissues, but in both cases they should exhibit a pronounced compatibility with the biological environment [1]. Most of the commonly employed materials (polymers, ceramics, metals, etc.) are non-toxic, have sufficient mechanical stability and elasticity, and are stable towards enzymatic degradation. However their non-physiological character often leads to undesirable processes, such as graft rejection, inflammations, infections, local tissue wasting and implant encapsulation, as well as thrombosis and embolization [2]. These biological responses are associated with a poor integration after implantation into the living body; that is, an inadequate interaction between the artificial material and the biological tissue. Controlled cell–biomaterial interaction is of utmost importance to avoid graft rejection and to favor a successful implantation. In particular, a strong mechanical contact between the implant surface and the surrounding tissue is required for the osseointegration of bone implants. Biofunctionalization of the implant material for rapid and specific cell colonization of their surfaces is of growing interest in implant technology and tissue engineering [3]. Biomimetic surface modification takes advantage of the power of specific biomolecular recognition events to control implant–tissue interactions without compromising the desirable bulk characteristics of an implant material. During the past few decades, those working in the fields of material science, surface engineering, chemistry, physics, biology, biochemistry and medicine have attempted to functionalize the surfaces of implant materials with bioactive molecules in order to enable signaling to adjacent cells and to obtain a desired cellular response. Cell–biomaterial interactions can be either specific or unspecific. Unspecific interactions are difficult to control because they are based on properties common to multiple cell types. The design of biomimetic materials takes profit of the specific interactions, related to defined chemical structures, such as ligands that interact with their corresponding cell surface receptor [4]. The most often-employed procedure to enhance cell adhesion and proliferation on synthetic surfaces targets the integrin receptors. 8.1.2 Integrins and RGD Sequence

The integrin family represents the most numerous and versatile group of cell adhesion receptors, which regulate the cell–cell and cell–extracellular matrix (ECM) interactions in multicellular organisms [5]. These interactions influence many fundamental cellular functions such as motility, proliferation, differentiation, and apoptosis. Therefore, integrins not only play a major role as anchoring molecules but they are also involved in many biological processes, such as embryogenesis, blood coagulation, immune response, hemostasis and regulation of the balance between cellular proliferation and apoptosis [6]. Integrins are heterodimeric transmembrane proteins composed of two noncovalently associated subunits (a and b). The 18 a and eight b known subunits combine to form 24 different heterodimers (Fig. 8.1) which differ in their ligand specificity [7].

8.1 Introduction

Fig. 8.1 The integrin family: the 24 known heterodimers.

The tripeptide sequence Arg-Gly-Asp (RGD) was identified as a minimal essential cell adhesion peptide sequence in fibronectin [8], since when cell-adhesive RGD motifs have been identified in many other ECM proteins, including vitronectin, fibrinogen, collagen, laminin, and osteopontin [9]. About half of the 24 integrins have been shown to bind to ECM molecules in a RGD-dependent manner [9]. The RGD sequence is the most effective and often-employed sequence to stimulate cell adhesion on synthetic surfaces. This is based on its widespread distribution and use in the organism, its ability to address more than one cell adhesion receptor, and its biological impact on cell anchoring, behavior, and survival. 8.1.3 Natural Proteins or Synthetic Peptides as Cell-Adhesive Molecules?

In the early studies, the surface of implant materials was coated with celladhesive ECM natural proteins which contain the RGD sequence in its structure [10]. However, the use of these proteins bears some disadvantages (Table 8.1) which prevent their practical use for medical applications. Most of these problems can be overcome when these macromolecular ligands are reduced to small recognition sequences, as small synthetic RGD peptides [11]. The RGD peptide sequence, structure and conformation play a crucial function in the ligand–receptor interaction and/or in the stability of the interaction. Sometimes, linear peptides experience slow enzymatic degradation [11a, 12], but small cyclic peptides are known to exhibit excellent long-term stability [13] as well as higher selectivity. Cyclic derivatives can interact with integrins more effectively than linear RGD peptides because the cyclization induces conformational stability as well as enhancing the preferred three-dimensional (3-D) structure for receptor interactions. However, cyclization usually results in low (or a lack of ) activity, and

111

112

8 Stimulation of Bone Growth on Implants by Integrin Ligands Table 8.1 Evolution, advantages and disadvantages of the different coating methods.

Integrin ligand

Advantages

Disadvantages

Proteins (1st generation)

Specificity, analogy to natural adhesion molecules

Enzymatic instability, immunogenicity, infection and inflammation risks, difficulties in anchoring, high costs

Peptides (2nd generation)

No risk of contamination, not immunogenic, higher temperature and pH stability, can be packed with higher density on surfaces, cost effectiveness

Enzymatic instability, lack of selectivity (e.g., osteoblast vs. platelets)

Cyclic peptides (3rd generation)

High selectivity, enzymatic stability

Moderate costs

Peptidomimetics (4th generation)

Very low costs, may be highly specific, high stability in the sterilization process as well as in vivo

–

only when the bioactive conformation is matched can super-activity and receptorsubtype selectivity be found. In the case of RGD peptides, a ‘‘spatial screening’’ procedure was applied to optimize the structure–activity relationship [14]. Modification of the amino acid sequences flanking the RGD motif or changing its 3-D structure has been shown to modify the ligand selectivity [15]. For example, the D-phenylalanine residue (f ¼ D-Phe) following the RGD binding sequence in peptides of the type cyclo(-RGDfK-) is essential to enhance the av -selectivity versus the platelet receptor aIIb b 3 (to induce the preferred adhesion of osteoblasts rather than of platelets). Cyclization also causes proteolytic stability. Both factors are of great importance for achieving a correct implant osseointegration in vivo. More recently, research into the design of specific integrin antagonists has focused on the development of non-peptidic integrin-selective mimetics, due to their specially desirable properties [16]. 8.1.4 Integrin-Mediated Cell Adhesion

The process of integrin-mediated cell adhesion involves four different, partly overlapped events [17]: cell attachment; cell spreading; organization of actin cytoskeleton; and the formation of focal adhesions. In the first step – the ‘‘initial’’ attachment – the cell contacts the surface and some ligand binding occurs; this allows the cell to resist gentle shear forces. Afterwards, the cell begins to flatten

8.1 Introduction

and its plasma membrane spreads over the substratum. The third effect noticed is the actin organization into microfilament bundles forming an actin cytoskeleton, sometimes referred to as ‘‘stress fibers’’. In the fourth step, the formation of focal adhesions occurs, which link molecules of the ECM to components of the cell’s actin cytoskeleton. During these four steps, integrins mediate physical anchoring and transmembrane signaling processes [5, 7]. The structure of the integrins allows them to function as bidirectional cellular signal transducers [18]. Conformational changes induced by the binding of ligands to integrins invoke signaling cascades inside the cell that regulate gene expression, activate kinases, and direct cytoskeletal organization (outside-in signaling) [5]. Alternatively, internal cellular activation can produce both conformational changes and multimeric clustering of the integrins, which results in non-consecutive binding to ligands, ECM components, as well as other cells (inside-out signaling) [19]. The type and degree of the signaling event is determined by the conformation and nature of the ligand, and is regulated by divalent cations bound to metal ion-dependent adhesion sites (MIDAS) on the integrin receptors. A great number of signaling events following the formation of focal adhesions are known. These include activation of focal adhesion kinase (FAK), extracellular signal-regulated kinase (ERK), small Rho GTPases and phosphatidyl inositol 4phosphate 5-kinase (PIP 5-kinase), and some elements of the mitogen-activated protein kinase (MAPK) pathway [5, 20], although many details are not yet clear. Nevertheless, it is well established that integrin-mediated cell spreading and focal adhesion formation induces survival and the proliferation of anchoragedependent cells [21]. The biofunctionalization of implant surfaces in order to mimic the biological environment and to stimulate a specific cell colonization is based on this effect (Fig. 8.2A). In contrast, loss of attachment causes apoptosis in many cell types, a situation referred to as ‘‘anoikis’’ [22]. Anoikis can even be induced in the presence of immobilized ECM molecules when non-immobilized soluble ligands such as RGD peptides are added (Fig. 8.2B). Based on this princi-

Fig. 8.2 The opposing effects of integrin ligands. (A) Immobilized ligands: these act as agonists of the extracellular matrix (ECM), leading to cell adhesion and survival. (B) Non-immobilized ligands: these act as antagonists of the ECM, leading to cell detachment and apoptosis.

113

114

8 Stimulation of Bone Growth on Implants by Integrin Ligands

Fig. 8.3 Scanning electron micrographs of adherent cells on substrates containing covalently grafted peptide (GRGDY). (A) Spheroid cells with no filapodial extensions. (B) Spheroid cells with one to two filapodial extensions. (C) Spheroid cells with more than two filapodial extensions. (D) Flattened morphology representative of well-spread cells. (Figure in according to Ref. [26])

ple, inhibition assays as well as some therapeutic applications of RGD peptides and peptidomimetics in the field of osteoporosis, renal failure, cancer and angiogenesis have been developed [23]. Recently, the role of the anoikis model was questioned concerning RGD peptide-induced apoptosis and the inhibition of angiogenesis [22b, 24]. Cell proliferation and apoptosis are two contrasting biological processes, both of which are integrin-dependent. Round cells are related to apoptosis, whereas cell spreading is related to its survival, focal adhesion formation, and proliferation [25]. Four types of adherent cell morphologies on surfaces have been described [26] (Fig. 8.3). With increasing concentrations of the integrin receptors on the surface (0.1, 1, 10 and 100 fmol cm2 ), the predominant morphology type shifts from A to D. Therefore, cell adhesion on biomaterials promoted by integrin ligands clearly depends on its surface density. Cell attachment as a function of RGD concentration shows a sigmoidal increase [27], indicating that there is a critical minimum density for cell response, which depends on the surface material and the type of cells. For comparable cell adhesion studies, it is important to know the density of the RGD peptides on the surface. The surface-bound RGD peptide may be determined using integrin-specific antibodies, although on the other hand an efficient method by radiolabeling with 125 I of a tyrosine-containing RGD peptide was recently reported [28]. The amount of bound peptide was seen to depend on the material and the concentration of the peptide in the coating solution used, although this was in the pmol cm2 range for all surfaces examined [poly(methyl methacrylate) (PMMA), titanium and silicon]. Nanoscale patterned surfaces were recently designed to address the precise molecular topology of focal adhesions by using block-copolymer micelle nanolithography [29]. These systems consist of a hexagonally packed rigid template of celladhesive gold nanodots coated with cyclo(-RGDfK-)-thiol peptide and separated by non-adhesive regions. The gold dots are small enough (

Handbook of Biomineralization - Medical and Clinical Aspects

Handbook of Biomineralization: Medical and Clinical Aspects (v. 3)

Biomineralization: Medical Aspects of Solubility