Department of Medicine, Helsinki University Central HospitalORTON Orthopaedic Hospital and Orton Foundation, Helsinki

Institute of Biomedicine/ Anatomy, University of HelsinkiDepartment of Orthopaedics and Traumatology, Helsinki University Central Hospital

The National Graduate School of Clinical InvestigationDepartment of Surgery, Mikkeli Central Hospital

OSTEOGENETIC GROWTH FACTORS AROUND LOOSENED HIP REPLACEMENTS

Ville Waris

ACADEMIC DISSERTATION

To be presented by permission of the Faculty of Medicine of the University of Helsinki for public discussion in the main auditorium of Mikkeli Central Hospital on January 24th, 2014,

at 12 o’clock noon.

Supervised by:

Professor Yrjö T. Konttinen, MD, PhDInstitute of Clinical Medicine, Department of Medicine, Biomedicum Helsinki 1, P.O. Box 700 (Haartmaninkatu 8), FI-00029 HUS, Finland.

Docent Dan Nordström, MD, PhDInsitute of Clinical Medicine, Department of Medicine, Biomedicum Helsinki 1, P.O. Box 700 (Haartmaninkatu 8), FI-00029 HUS, Finland.

Reviewed by:

Docent Petri Virolainen, MD, PhDDepartment of Orthopaedics and Traumatology, Surgical Hospital, Turku University Hospital, P.O. Box 28, FI-20701, Turku, Finland.

Docent Nils Hailer, Dr. med.Department of Orthopedics, Institute of Surgical Sciences, Uppsala University Hospital, SE-75185 Uppsala, Sweden.

Opponent:Professor Heikki Kröger, MD, PhDDepartment of Orthopedics and Traumatology, Kuopio University Hospital, Kuopio, Finland; Bone and Cartilage Research Unit (BCRU), University Of Eastern Finland, Kuopio, Finland.

Tieteellinen Tutkimus ORTONin julkaisusarja, A: 36Publications of the ORTON Research Institute, A: 36P.O. Box 29, 00281 Helsinki

ISBN 978-952-9657-72-8 (paperback)ISBN 978-952-9657-73-5 (pdf)ISSN 1455-1330http://ethesis.helsinki.fi

Helsinki University Printing House

Helsinki 2013

CONTENTS

List of original publications .................................................................................................. 5

Abbreviations .......................................................................................................................... 6

Abstract ....................................................................................................................................8

1 Introduction .....................................................................................................................10

2 Review of the literature ................................................................................................. 13 Bone tissue and bone cells ...................................................................................13 Cytokines and their regulatory role in bone metabolism ..................................17 Osteoarthritis .......................................................................................................20 Total hip replacement ..........................................................................................21 Aseptic loosening ................................................................................................. 23 Bone formation enhancing growth factors ........................................................ 32 Basic Fibroblast Growth Factor ................................................................. 32 Transforming Growth Factor-βs ............................................................... 34 Platelet Derived Growth Factor ................................................................. 37 Insulin-like Growth Factors ....................................................................... 39 Bone Morphogenetic Proteins ....................................................................41 Vascular Endothelial Growth Factor ......................................................... 43

3 Aims of the study ...........................................................................................................46

4 Materials and methods ..................................................................................................48 Patients and samples ...........................................................................................48 Immunohistochemistryandimmunofluorescencestaining ...........................50 Cell cultures and stimulations ............................................................................ 54 Reverse transcription polymerase chain reaction, RT-PCR ............................ 55 Quantitative real-time reverse transcription polymerase chain reaction, QRT-PCR .................................................................................... 55 Quantificationofimmunohistochemicalstaining ............................................ 56 Statistical analysis .................................................................................................57

5 Results and discussion ...............................................................................................................58 Histopathological evaluation of interface tissue and pseudocapsular tissue ..................................................................................58 Wear particulate debris ....................................................................................... 59 Basic FGF in aseptic loosening .................................................................. 59 TGF-β1andTGF-β2inasepticloosening ..................................................61 PDGF in aseptic loosening ......................................................................... 63 IGF-I and II in aseptic loosening ............................................................... 65 BMPs in aseptic loosening .........................................................................66 VEGF and its receptors 1 and 2 in aseptic loosening ...............................68 Patient and control sample selection .........................................................71 Relevance of methods ................................................................................. 72 Possible roles of the studied growth factors in aseptic loosening ........... 73

6 Conclusions .....................................................................................................................................76

7 Acknowledgements ....................................................................................................................77

References ..............................................................................................................................................79

Original publications ......................................................................................................................104

5

LIST OF ORIGINAL PUbLICATIONS

I Waris V, Xu JW, Nordsletten L, Sorsa T, Santavirta S, Konttinen YT: Basic fibroblastgrowthfactor(bFGF)inthesynovial-likemembranearoundloosetotalhip prosthesis. Scand J Rheumatol. 1996 25:257-262.

II Konttinen YT, Waris V, Xu JW, Jiranek WA, Sorsa T, Virtanen I, Santavirta S: Transforming growth factor-ß (TGF-ß) 1 and 2 in the synovial-like membrane between the implant or cement to bone interface in loosening of total hip replacement (THR). J Rheumatol. 1997 24:694-701.

III Xu JW, Konttinen YT, Li TF, Waris V, Lassus J, Matucci-Cerinic M, Sorsa T, Santavirta S: Production of platelet-derived growth factor in aseptic loosening of total hip replacement. Rheumatol Int. 1998 17:215-221.

IV Waris V, Zhao DS, Leminen H, Santavirta S, Takagi M, Nordsletten L, Konttinen YT: Insulin-like growth factors I and II in the aseptic loosening of total hip implants. Scand J Rheumatol. 2004 33:1 –5.

V Waris V, Waris E, Sillat T, Konttinen YT: BMPs in periprosthetic tissues around aseptically loosened total hip implants. Acta Orthop Scand. 2010 81: 420–426.

VI Waris V, Sillat T , Waris E, Virkki L, Mandelin J, Takagi M, Konttinen YT: Role and regulation of VEGF and its receptors 1 and 2 in the aseptic loosening of total hip implants. J Orthop Res. 2012 30:1830-6.

6

AbbREVIATIONS

ABC avidin-biotin-peroxidase complexAP alkaline phosphataseAPAAP alkaline phosphatase anti-alkaline phosphatasebFGF basicfibroblastgrowthfactorBMP bone morphogenetic proteinBMPR bone morphogenetic protein receptorBSA bovine serum albuminCDH congenital dislocation of hipCM-IFM cultured media from interface membraneCoCrMo cobalt-chromium-molybdenum alloyDNA deoxyribonucleic acidDMEM Dulbecco’smodifiedEagle’smediumEC endothelial cellECM extracellular matrixGH growth hormoneIg immunoglobulinIGF insulin-like growth factorIL interleukinIT interface tissueMAb monoclonal antibodyMMP matrix metalloproteinaseMSC mesenchymal stem cellmRNA messenger ribonucleic acidOA osteoarthritisOB osteoblastOC osteoclastOCT optimal compound for tissue embeddingOPG osteoprotegerinPAI plasminogen activator inhibitorPBS phosphate buffered salinePDGF platelet derived growth factorPMMA polymethylmethacrylatePT periprosthetic tissue; pseudocapsule and interface membranesRA rheumatoid arthritisRANK receptoractivatorofnuclearfactorκBRANKL receptoractivatorofnuclearfactorκBligand

7

rTHR revision total hip replacementRT room temperatureRT- PCR reverse transcription polymerase chain reactionqRT-PCR quantitative real time reverse transcription polymerase chain reactionSEM standard error of meanTBS 50 mM Tris buffered 150 mM saline, pH 7.4TGF transforming growth factorTHA total hip arthroplastyTHR total hip replacementTiAlV titanium-aluminium-vanadium alloyTRAP tartrate resistant acid phosphataseUHMWPE ultra-high molecular weight polyethyleneuPA urokinase type plasminogen activatorVEGF vascular endothelial growth factor

8

AbSTRACT

Total hip replacement (THR) is an efficacious, reliable, and cost effectiveprocedureforreconstructionofhipjointsafflictedbyendstagearthritis.Asepticlooseningisoneofthemostsignificantlong-termcomplicationsofTHRandthemajor cause for clinical failure of the procedure. The aetiology is multifactorial. Both biological and mechanical factors seem to play important roles in aseptic loosening. The formation of a synovial-like membrane between prosthesis or cementandboneisacommonpathologicfindingduringrevisionTHRperformedfor aseptic loosening. The hypothesis in this study was that cytokines, which stimulate bone formation, are decreased in interface tissue (IT) in patients with loose THR compared to control synovial membranes.

In this study, the histopathological features of IT were found to be very similar to synovial tissue from OA. However, particulate debris and macrophages/foreign body giant cells were found in most ITs. Immunohistochemical staining showed up-regulated expression of bFGF, PDGF, TGF-ß1 and TGF-ß2 in interface and pseudocapsular tissues but down-regulated expression of IGF-I in the same areas. IGF-II was down-regulated in pseudocapsular tissues. Expression of BMP-2, -4, -6 and 7 as well as VEGF and its receptors 1 and 2 was similar in ITs when compared to control samples in general, but expression of BMP-4 and VEGF receptor 1 was up-regulated in IT macrophage-like cells and VEGF receptor 1 and2alsoinITfibroblast-likecells.Thecellsoforiginofthesecytokinesweremacrophages,fibroblasts,andendothelialcells.

This study clarified the cytokine spectrum found in interface and pseudocapsular tissue in loose THR. The results do not support the hypothesis that these cytokines, which are able to stimulate bone formation, would be consistently decreased in these tissues. All of the studied cytokines can induce scar tissue development and many of them also blood vessel formation, tissue inflammationandsomeevenenhanceosteoclasticboneresorption.Theymostprobablyhavemixedpositiveandnegativeeffectsonimplant-bonefixationinthe course of aseptic loosening of THR, depending on multiple factors like local concentration of different cytokines, cell-cell contact modulation, periprosthetic tissue (PT) hypoxia and pH, implant micromotion, and systemic factors such as hormones.

Despite technological improvements, such as the introduction of highly cross-linked UHMWPE, ceramic heads and cups or diamond-like coatings, it is unlikely that particulate debris and loosening can be completely eliminated. Therefore, additional studies are necessary to improve our understanding of

9

the complex cytokine network in these tissues in aseptic loosening of THR. Through further understanding of these aseptic loosening processes, it may be possible in the future to modulate or prevent cytokine expression or their effects immunologically or pharmacologically.

10

1 INTRODUCTION

The total hip replacement (THR) procedure has advanced steadily after the pioneering work of Sir John Charnley (1961). THR is now a well-accepted treatment ofpatientswithend-stagehipdisorderscausedbydegenerativeorinflammatorydiseases, aseptic vascular necrosis or hip dysplasia providing marked symptomatic relief and improved function. Approximately 835,000 primary and 125,000 revision THR operations are performed annually worldwide (Kurtz et al., 2010).

As improvements in prosthetic design and surgical technique have reduced the incidence of prosthetic breakage and infection, aseptic loosening is the major cause of failure of THR at the present time. Despite intense interest and study, the pathological mechanisms leading to bone resorption around the implant and eventual aseptic loosening are not fully understood. Factors that interact to produce aseptic looseningcanbeclassifiedintoseveralprocesses, includingmechanicalfactors, material properties of the implants, and biological and host factors.

Much of the research regarding THR currently focuses on biological mechanisms of aseptic loosening. Wear debris generated on articulating surfaces is phagocytosed by macrophages and can subsequently activate macrophages to release a variety of cytokines and proteolytic enzymes that stimulate osteoclastic bone resorption and degrade extracellular matrix to provoke a cascade leading to osteolysis, extension of the effective joint space and aseptic loosening (Howie et al., 1993; Maloney and Smith, 1995; Konttinen et al, 1997a; Goodman et al., 1998; Konttinen et al., 2005; Purdue et al., 2007; Tuan et al., 2008; Brown et al., 2009; Hallab et al., 2009). It seemsthatthemacrophageisthekeycellinthisprocess.Inaddition,fibroblastsand osteoblasts may also be actively involved in the pathogenesis of loosening via formationoffibrous implantcapsuledisturbingosseointegration(directcontactbetween implant/cement and bone) and via diminished de novo bone formation around the implant (with net loss of bone when the rate of the osteolytic process exceeds that of the formation of new bone).

The cytokine research in aseptic loosening has shifted during the course of this study towards the mechanisms of macrophage and osteoclast activation, function andregulationthroughthereceptoractivatorofnuclearfactorκBligand(RANKL)-system (Mandelin et al., 2003; Tuan et al., 2008; Wang et al., 2013).

Bone is continuously remodelled through combined and usually balanced bone resorption and bone formation. Bone remodelling is regulated by systemic hormones, local factors and cytokines. Local factors may be more important than systemic hormones for the initiation of physiologic bone resorption and normal bone remodelling sequence to satisfy the local biomechanical needs. Because bone

11

remodelling occurs in discrete and distinct packets throughout the skeleton, it seems probable that the cellular events are controlled by factors generated in the microenvironment of bone. Cytokines as local factors have important effects on normal and diseased bone as regulators of cell behaviour.

In the remodelling of normal adult bone, resorption of old or injured bone and formation of new bone are balanced and coupled together and there is a balance between osteoclast stimulating/osteolytic and osteoblast stimulating/osteogenetic cytokines. In most diseases of the skeleton there is an increased remodelling of the bone, in which osteoclastic resorption precedes and exceeds osteoblastic bone formation, resulting in systemic (osteopenia, osteoporosis, osteomalacia) or local (osteolytic lesions) bone loss or impairment.

Our hypothesis is that particulate debris produced as a result of wear of joint prosthesesinducesachronicinflammatoryresponseintheperiprosthetictissuesandthatcytokines(theinflammatoryfunctiolaesa)secretedbyactivatedmacrophagesplay an important role in the development of osteolysis and aseptic loosening. Thechronicforeign-bodytypeinflammatoryreactioninthesynovialmembrane-like interface tissue (IT) seems to activate predominantly synthesis of osteoclast stimulating, ‘osteolytic’ cytokines and too little, or at inappropriate sites, of those able to stimulate osteoblasts and new bone formation.

The aim of the present study was to investigate the presence, cellular localization and extent of expression of osteoblast stimulating cytokines in the IT between bone, cement and implants and in pseudocapsular tissues of aseptically loosened prosthetic hip joints.

THRisanefficacious,reliable,andcosteffectiveprocedureforreconstructionofhipjointsafflictedbyendstagearthritis.Asepticlooseningisoneofthemostsignificant long-term complications of THR and for the clinical failure of theprocedure. The aetiology is multifactorial. Both biological and mechanical factors seem to play important roles in aseptic loosening. Instead of osseointegration and implant capsule formation, an increase in the volume of the effective joint space and formation of a synovial-like membrane between prosthesis or cement and boneisacommonpathologicalfindingfoundduringrevisionTHRperformedforaseptic loosening. The hypothesis in this study was that bone formation stimulating cytokines are decreased in ITs in patients with loose THR compared to control synovial membranes.

In this study, the histopathological features of IT were found to be very similar to synovial tissue from OA. However, particulate debris and macrophages/foreign body giant cells were also found in ITs. Immunohistochemical staining showed upregulatedexpressionofbFGF(basicfibroblastgrowthfactor),TGF-ß1andTGF-ß2(transforming growth factors) and PDGF (platelet derived growth factor). IGF-I (insulin like growth factor I) was down-regulated in IT and pseudocapsular tissue.

12

IGF-II was down-regulated in pseudocapsular tissue. IGF-II was up-regulated in pseudocapsular tissue. Expression of BMP-2, -4, -6 and 7 (bone morphogenetic proteins) as well as VEGF (vascular endothelial growth factor) and its receptors 1 and 2 was similar in ITs when compared to control samples in general, but expression of BMP-4 and VEGF receptor 1 was up-regulated in IT macrophage-like cells and VEGFreceptor1and2alsoinITfibroblast-likecells.Thecellsoforiginofthesecytokinesweremacrophages,fibroblasts,andendothelialcells(EC).

All the studied cytokines share an ability to stimulate bone formation. Their spectruminITandpseudocapsulartissueinlooseTHRwasclarifiedinthisstudy.The results do not support the hypothesis that the studied cytokines would all be decreased in these tissues. In addition to their bone formation stimulating properties, all of the studied cytokines can also induce scar tissue development andmanyof themalsobloodvessel formation, tissue inflammationandsomeeven enhance osteoclastic bone resorption. They most probably have positive and negativeeffectsonimplant-bonefixationinthecourseofasepticlooseningofTHR,depending on multiple factors like local concentration of different cytokines, cell-cell contact modulation, periprosthetic tissue hypoxia and pH, implant micromotion, and systemic factors such as hormones.

Despite technological improvements, such as the introduction of highly cross-linked, radical free UHMWPE, ceramic heads and cups or diamond-like coatings, it is unlikely that particulate debris and aseptic loosening can be completely eliminated. Therefore, additional studies are necessary to improve our understanding of the complex cytokine network in these tissues in aseptic loosening of THR. Through further understanding of these aseptic loosening processes, it may be possible in the future to modulate or prevent cytokine expression or their effects immunologically or pharmacologically.

13

2 REVIEW OF THE LITERATURE

Bone tissue and Bone cells

Bone is a specialized connective tissue that serves four main functions: 1) mechanical: support and location of muscle attachment for locomotion; 2) protection of vital organs and bone marrow; 3) metabolic: as a reserve of ions, especially calcium (Ca) and phosphate (PO4), for the maintenance of serum homeostasis, which is essential to life (Baron, 1996) and 4) cellular: by providing niches for the formation of haematopoietic cells and bone marrow derived mesenchymal stem (stromal) cells (MSC). Bone tissue contains various types of dynamic structural cells, of which the bone-forming osteoblast (OB) and bone-resorbing osteoclast (OC) are mainly responsible for bone remodelling (Figure 1.).

OBs are thought to be derived from undifferentiated MSCs, which further differentiateintoosteocyteswhenembeddedincalcifiedbonetissueortoboneliningcells (Bonewald, 2011). The external bone surface is covered by richly innervated periosteum, the deeper layer of which contains the Cambium (osteogenic) cell layerasanimportantsourceofMSCsforfracturerepair,protectedbyafibrousperiosteum. Similarly, the inner surface of the bone is covered by endosteum (Dwek, 2010). Actually, 90-95% of human bone cells are osteocytes, which can live decades (Bonewald, 2011). Osteocytes have long and slender cellular dendritic processes locatedinbonecanaliculibathedbyperiosteocyticinterstitial(canalicular)fluid.Processes form a network through which osteocytes can communicate with each other, with OBs and with terminally differentiated OBs on bone surfaces, the above mentionedboneliningcells.Withthehelpofthefluidshearstressandgapjunctions,this intercommunicating cellular network may be an important stress and damage sensor (Bonewald, 2011).

OCs are multinucleated cells present only in bone. It is believed that OC progenitors are recruited from haematopoietic tissues, formed in and released from the bone marrow into blood and recruited then from blood back to bone marrow, regulated by chemorepulsive and chemokinetic stimuli (Ishii et al., 2011; Yu et al., 2003; Koizumi et al., 2009). OC and also foreign body giant cell progenitors then proliferate and differentiate into mononuclear preosteoclasts and fuse with each other to form multinucleated OCs or, alternatively, to foreign body giant cells. Osteoclasts haveauniquemorphologyandfunctionbeingabletoresorbcalcifiedbonebymakingHowship’s lacunae and resorption pits. Recently discovered CD68 positive tissue macrophages in bone, osteomacs, reside on bone surfaces between or on top of bone lining cells. Osteomacs cover the sites where bone remodelling takes place as a

14

canopy, forming a protective bone remodelling compartment or microenvironment and probably regulate the process (Pettit et al., 2008; Raggatt et al., 2010).

Boneextracellularmatrix (ECM) containsfibrinous components: collagens(main component of the ECM) and proteoglycans. It also contains more than 160 non-collagenous,non-fibrinousproteinsandpolypeptides:forexampleγ-carboxyglycoproteins,osteonectin,fibronectin, thrombospondin,bonesialoproteinandosteopontin. There are cross-links between the collagen and the proteoglycan network and/or attachment sites for various bone cells, and enzymes like alkaline phosphatase and matrix metalloproteinases (MMP) participating in the formation and remodelling of bone and bone-related soft tissues. Minerals, mainly impure hydroxyapatite ([Ca10[PO4]6][OH]2), are precipitated in newly synthesized osteoid bone ECM in a mineralization process requiring collagen and active 1,25(OH)2-vitamine D3 and regulated by hormones, like sex steroids, and by osteocytes that have been embedded in osteoid (Bonewald, 2011). ECM molecules play critical roles in bone formation and remodelling. ECM assists in bone cell mobilization, adhesion, mechanochemical regulation, differentiation, proliferation, growth, functional change and apoptosis. Furthermore, ECM sequester and modulate growth factors enabling their storage in bone tissue as well as their release and activation during the so called coupling of OC and OB function (Allori et al., 2008). Other molecules supposed to play a role in coupling (repression of OCs and stimulation of OBs at the reversal) include sphingosine-1-phosphate, Wnt10b and BMP-6 produced and secreted by mature osteoclasts, leading to recruitment and/or differentiation of osteoblast precursors/MSCs (Pederson et al., 2008), and bidirectional signalling between Ephrin-B2 on osteoclast surface and EphB4 tyrosine kinase on osteoblast surface (Zhao et al., 2006).

When bone remodelling is not needed, osteocytes secrete sclerostin (van Bezooijen et al., 2004), which inhibits the stimulation of bone lining cells and initiation of bone remodelling. At the same time, osteocytes secrete TGF-ß, which inhibits osteoclastogenesis (Heino et al., 2002). Osteocytes are able to sense mechanicalstimuli,whichproduceinterstitialfluidshiftsincanalicularnetworkand turn them into chemical signals involving cyclo-oxygenase 2, prostaglandin E2 (PGE2), various kinases, Runx2, and nitrous oxide (Clarke, 2008; Allori et al., 2008). Parathyroid hormone and local bone overloading causes cyclic bone deformation, which inhibits sclerostin secretion by osteocytes. This leads to apposition of new bone. The resulting thicker bone resists bone deformation and osteocyte sclerostin secretion starts again completing the negative feedback loop (Bonewald, 2011).

Bone is remodelled continuously during adulthood as OCs resorb old bone and OBs subsequently form new bone in a in a basic multicellular unit (BMU, see Figure 1). These two closely coupled events are responsible for renewing the skeleton while maintaining its anatomical and structural integrity.

15

Figure 1. BMU and bone remodelling (modified from Raggatt et al., 2010).

Under normal conditions, bone remodelling proceeds in activation-reversal-formation cycles (Allori et al., 2008; Raggatt et al., 2010). Remodelling sites are thought to develop mainly randomly but can also be targeted to areas that require repair as signalled by osteocyte apoptosis (Eriksen, 2010).

Bone lining cells sense the signals mentioned above and recruit OC precursors to the remodelling site by secreting chemokines, which also stimulate cellular fusion and expression of tartrate-resistant acid phosphatise (TRAP) and calcitonin receptor, but cannot alone activate the bone resorbing function. Bone lining cells, OBs and other local cells produce macrophage colony-stimulating factor (M-CSF) andreceptoractivatorofnuclearfactorκBligand(RANKL),whichtogethermaintainand induce OC precursors and later assist in OC differentiation, activation and function. At the same time, the production of RANKL inhibiting osteoprotegerin (OPG) by the same cells is reduced (Allori et al., 2008; Raggatt et al., 2010), which

16

tilts the RANKL/OPG balance in favour for osteolysis. Apoptotic and necrotic osteocytes are also able to recruit and activate OCs through the RANK/RANKL/OPG system in sites of bone damage (Bonewald, 2011). Local cytokines and systemic hormones regulate the RANK/RANKL/OPG system in a complex way (see later descriptions). Megakaryocytes are located in bone marrow and produce platelets. Interestingly, in vitro megakaryocytes have been found to express RANKL, OPG and a yet unknown soluble anti-osteoclastic factor as well as to enhance OB proliferation and differentiation (Lorenzo et al., 2008).

OBs produce MMPs, which digest bone surface and uncover anchoring sites where activated OCs can adhere. Subsequently, OCs form a sealed erosion cavity called Howship's lacuna between the OC and the bone surface. Proton pumps acidify the lacuna and hydrochlorid acid dissolves minerals, followed by collagenolysis mediated by cathepsin K and acidic endoproteinases. At the same time OC is activated to transcytotic transport of bone pieces from the Howship’s lacuna to the apical OC domain and to release them there for further processing mainly by macrophages (Salo et al., 1997; Caetano-Lopes 2011). Shortly after the OCs have left the resorption site its surface is processed by reversal cells which are mononuclear cells or derived from osteoblast lineage. They form a cement line, marking the outer border of the osteon in cortical compact and hemiosteon in trabecular porous bone, before OBs invade the area and begin to build new bone. Osteon refers to the basic structural unit of bone. The bone defect formed by the osteoclasts in thecuttingcone isfilledwithosteoid (amatrixof collagenandothernon-collagenous proteins) in the closing cone. This organic bone matrix is eventually mineralised in a process which can be measured in bone biopsies using dynamic bone histomorphometry. Bone surface becomes again covered by bone lining cells after the bone formation has ceased. Some of the OBs have become embedded in the bone matrix in lacunae and have thus become networking osteocytes. At, or shortly after osteoid mineralization the osteocytes start again to secrete sclerostin. ThisinhibitsoverfillingoftheBMUwithnewboneandkeepsboneliningcellsquiescent, which inhibits further remodelling. Exact mechanisms of mineralization and termination of the remodelling cycle are not clear (Clarke, 2008; Moester et al., 2010; Raggatt et al., 2010).

Bone remodelling is a coupled process in which bone resorption is normally followed by new bone formation. The coupling is not fully understood, but probably involvesreleaseofbone-boundanabolicgrowthfactors(e.g.TGF-β,IGF-I,IGF-II,BMPs, PDGF and bFGF) during osteoclastic resorption. These factors are activated (released from their storage sites) by the acidic pH. Coupling factors released or expressed by OCs, strain gradient in Howship’s lacuna and other coupling mechanisms have also been proposed (Martin and Sims, 2005; Clarke, 2008; Caetano-Lopez, 2011). Osteomacs cover the resorption site where they are in close

17

proximity with OCs and OBs. They have a broad range of effects on OCs and OBs and therefore are also important coupling candidates (Pettit et al., 2008).

Bone remodelling is regulated by systemic hormones and local factors which affect cells of OC or OB (including osteocytes and lining cells) lineage and exert their effects on 1) the replication of undifferentiated cells, 2) the recruitment of cells, 3) the differentiation of cells and 4) the differentiated function of cells. The regulatory mechanisms are complex and incompletely understood. Systemic hormones modulate this activity throughout the skeleton. Parathyroid hormone (PTH), calcitonin and active vitamin D3 affect levels of serum calcium, partially throughtheir influenceonbonecells.Otherhormonesthat influencebonecellfunction include thyroxin, glucocorticosteroids, androgens and estrogens. Systemic signals like estrogen deficiency and corticosteroid treatment can also induceosteocyte apoptosis and activate BMUs (Clarke, 2008).

Local factors such as cytokines may have the most direct effect on bone remodelling. Reduced bone loading leads to increased osteocyte apoptosis and thus to local TGF-ß depletion initiating osteoclastogenesis and the bone remodelling cycle through activation of RANK/RANKL/OPG –system (Allori et al., 2008; Raggatt et al., 2010). The role of nerves innervating bone is incompletely understood but may also play an important role in bone remodelling (Konttinen et al., 1996a).

cytokines and their regulatory role in Bone metaBolism

Cytokines are important as local mediators of intercellular communication in normal or diseased tissues. Immunologists used previously the term “cytokine” to describe molecules which are released by macrophages or lymphocytes and modulate cellular activities. Cell biologists used the term “growth factor” for peptides and proteins which stimulate or inhibit cell division or differentiation. Cytokines and growth factors are similar in their general biological actions and mechanisms. For example, they are secreted by cells and modulate the activity of these or other cells. The term cytokineislessrestrictiveandmoreaccuratelyreflectsthebiologicactivitiesofthesefactorsandincludesalsoproteinsoriginallyclassifiedasgrowthanddifferentiationfactors (Nathan and Sporn, 1991).

Mechanisms of cytokine influence on cellular activities

Intracrine effect: cytokines can function intracellularly in the cell that synthesized themwithoutsurfaceexpressionorsecretiontotheextracellularfluid.Theycan

18

be released by the cytokine producing cells in microvesicles, which are internalized by target cells, in which the microvesicles fuse with the extracellular membranes of those cells. Intracrine cytokine effects have been described by Re and Cook (2009) in a review article with references for intracrine action of bFGF, IGF-I, PDGF, TGF-ß and VEGF.

Autocrine effect: a mechanism involving the synthesis of a cytokine by a cell thatinfluencesthegrowthandfunctionofthesamecellviaaspecificreceptoronits surface. For example, autocrine BMP is required for OB differentiation (van der Horst et al., 2002).

Juxtacrine effect: thismechanism involves specific cell-to-cell contacts. Itoccurs via an interaction of a membrane-bound form of a cytokine with a receptor on a surface of an adjacent cell. Another mode of juxtacrine interaction involves extracellular matrix-associated (rather than membrane-anchored) forms of a cytokine produced by one cell and affecting the neighbouring cell via the juxtacrine matrix. VEGF produced by microvessel pericytes are thought to assist in EC survival in a juxtacrine manner (Darland et al., 2003).

Paracrineeffect:synthesisofacytokinebyonecellthatinfluencesthegrowthandfunctional activities of nearby cells expressing the receptor for the secreted cytokine. Within tissues, cytokines exert their effects most often in a paracrine fashion. For example, VEGF can have a paracrine effect in the communication between ECs and OBs as it couples angiogenesis with bone formation (Clarkin et al., 2008).

Endocrine effect: Cytokines can also have endocrine actions. As an example, circulating IGF-I produced by liver can enhance skeletal tissue growth (Ohlsson et al., 1998).

Cytokines can act via trans-signalling. IL-6 can bind to IL-6 receptor A, forming a complex that can then activate interleukin-6 receptor B positive, but IL-6 receptor A negative cells. IL-6 receptor A is the ligand binding component and IL-6 receptor B is the signalling component of the complete IL-6 receptor. Via trans-signalling soluble IL-6/IL-6 receptor A complex can regulate cells which cannot directly bind IL-6 (Youinou and Jamin 2009). Cytokines may also act via trans-presentation. Interleukin-15bindswithhighaffinitytointracellularIL-15receptorA.ThisIL-15/interleukin-15 receptor A complex is translocated to the cell surface and there IL-15 receptor A-bound IL-15 can stimulate neighbouring IL-15 receptor beta and common gamma (gamma C) chain positive cells (Castillo and Schluns, 2012).

Based on their cellular origin and principal biologic activities, cytokine families can be divided into:

1. interleukins2. tumour necrosis factors3. growth factors

19

4. colony stimulating factors5. interferons6. chemokines

An important general principle is that each cytokine exhibits multiple effects, opposing and antagonizing or supplementing and synergizing with other cytokines in a context dependent manner. Thus, a biologic response observed in cell culture in vitro and especially in tissue in vivo represents the net effects of multiple factors. Cytokinesbindtospecificreceptorsontargetcellsandinduceintracellularsignaltransduction. The resultant biologic responses to soluble, membrane-bound or receptor-associated intracellular cytokines are caused by changes in gene transcription, messenger ribonucleic acid (mRNA) translation, and protein synthesis and secretion or by other means, such as by changes in protein kinase activities.

Cytokines play critical roles in mediating cellular interactions within skeletal tissues.Mostofthecytokineshaveeffectsonbone,buttheexistenceofbone-specificcytokines is still questioned. The regulatory functions of the cytokines are performed throughout life, beginning with embryogenesis, bone growth and development, in the mature organism with balanced bone remodelling, and in the aging skeleton with senile osteopenia and osteoporosis. Mechanisms by which cytokine networks regulate local tissue and cellular processes in bone can be listed as:

1. effects on cell proliferation, differentiation, and death2. up- or down-regulation of cell products, proteinases, other cytokines,

cytokine-binding proteins, and soluble receptors or small molecular weight molecules, such as prostaglandins, leukotrienes, and products of oxygen metabolism

3. effects on production of extracellular reactive oxygen and nitrous species4. modulation of expression of cell-surface constituents, including

receptors and adhesion molecules (Goldring and Goldring, 1996) and5. modulation of ion channels

In vitro studies seem to indicate that cytokines may be divided into those that cause primarily either resorption or formation of bone. Some cytokines have been reported to be clearly able to do both. For example, TGF- ß and PDGF in different concentrations and conditions can cause either resorption or formation (Konttinen et al., 1997a).

20

osteoarthritis

Osteoarthritis (OA) is the most common form of arthritis and one of the leading causes of pain and disability worldwide. It can be divided into primary or secondary OA and to a radiographic and/or clinically symptomatic OA, with the primary and/or radiographic forms representing the most common changes. Primary OA affects most commonly knee, hip and small hand joints. The symptoms include joint pain, tenderness, stiffness, locking and sometimes effusion. Prevalence of radiographic OA with symptoms in adults over age 50 varies from 5% (hip) to 13% (knee). OA has multiple risk factors: genetic factors, aging, female sex, obesity, high bone density, joint injury, occupational and recreational joint usage, reduced muscle strength, joint laxity and joint malalignment (National Collaborating Centre for Chronic Conditions, 2008).

Different types of joint injuries may trigger repair processes where all tissues of the joint tissues play a role, showing increased cell activity with tissue trauma, degradation and new tissue production. This results in a structurally altered joint. Sometimes either the original insult or the sum of continuing injuries is too great or repair potential is compromised. In those individuals the continuing tissue damage leads eventually to symptomatic OA. Accordingly, there is great variability in clinical presentation and outcome, both between individuals and at different joint sites.

Today,moreemphasishasbeengiventoinflammatoryandhostchangesinOA pathophysiology. OA is a metabolically active, dynamic process. It involves all joint tissues: in addition to cartilage, subchondral bone (Burr and Gallant 2012), synovium (Saito et al., 2002), joint capsule and ligaments (Madry et al., 2012) and even extra-articular tissues, such as the nervous system (e.g. gate disturbances and impairment of the efferent and afferent motor programs; Konttinen et al., 2006). Even regional muscles may atrophy. Main pathological changes include loss of articular cartilage and remodelling of subchondral bone (thickening and stiffening)withsynovialproliferation/inflammation(secondarysynovitis)andnew bone formation at the joint margins (osteophytes). Pain in OA is probably caused secondarily by osteoarthritic synovitis or perhaps in part by bone marrow oedema and increased intramedullary pressure (Felson et al., 2001, Kiaer et al. 1988). Synovial lining of all joints, tendon sheats and bursae is composed of macrophage-liketypeAliningcellsandfibroblast-liketypeBcells.Beneaththesynovial lining is a vascularized sublining, connective tissue stroma. Both areas producesensitizing,algogenicandpro-inflammatorymediatorsinOAsynovitisandsecretethemtosynovialfluid.Chondrocytesanddegeneratingcartilagematrixare also able to produce the mediators, such as endogenous damage-associated molecular patterns (alarmins) and may even start the whole pathologic cascade, withthechondrocyteastheprimaryinflammatorycell(Konttinenetal.,2012).

21

Foramorespecificpathobiologydescriptionsee thereviewbyAbramsonandAttur(2009).Briefly,OAjointtissuesproducenumerouschemokines,cytokines,prostaglandins, proteases, free radicals and nitrous oxide capable to mediate cartilage destruction, inflammation and disease progression. Local anabolic,tissue reparative cytokines/ growth factors like BMPs, IGF-I, bFGF and TGF-ß areproducedininsufficientamountstocounteractcartilagedestructionandtostimulate cartilage repair. The repair potential of cartilage tissue is poor due to the low cellularity (a low cell-to-matrix ratio) and lack of blood and lymphatic vessels and nerves. However, these cytokines probably take part in osteophyte formation and sclerosis of subchondral bone.

Recently, inflammatorycomplementcascadedysregulationinsynovial jointshas been shown to play a central role in the late phases of the pathogenesis of osteoarthritis (Wang Q et al., 2011). Cartilage oligomeric matrix protein (COMP) is astructuralcomponentofcartilage,whereitfacilitatescollagenfibrillogenesisandcross-linking.IthasgotanactiveroleinOAinflammationsinceitcanmodulatecomplementactivation.ElevatedamountsofCOMParefoundinsynovialfluidduring increased turnover of cartilage associated with active joint disease, such as rheumatoid arthritis (RA) and OA (Happonen et al., 2010).

Treatment of OA generally involves a combination of exercise, physical therapy, lifestylemodification,analgesics,oralglucosamineand intra-articularuse of corticosteroids and hyaluronan. So far, no drug has been proven to be chondroprotective or –reconstructive. If pain becomes debilitating, joint replacement surgery may be used to improve the quality of life by relieving pain and improving movement and functions.

total hip replacement

THR is now entering its sixth decade of clinical success since the introduction of polymethylmethacrylate (PMMA) cemented low-friction arthroplasty by Sir John Charnley (Charnley, 1961). THR has become well accepted throughout the world as a means of correcting a variety of hip disorders, including those caused by severe rheumatoid arthritis (RA), OA, severe traumatic arthritis, aseptic necrosis of the femoral head and congenital dysplasia or dislocation of the hip joint. Approximately 835,000 primary and 125,000 revision THR operations are performed worldwide with an annual growth rate of 3.2% (Kurtz et al., 2010).

Modern THRs consist of a femoral stem usually made of titanium or chromium-cobalt alloy, a separate head typically made of chromium-cobalt alloy and an acetabular cup component. This can be a monoblock type i.e. consisting of only chromium-cobalt alloy or ultra-high molecular weight polyethylene (UHMPWE)

22

coated with titanium, hydroxyapatite or tantalum. It can also be a modular type cup consisting of metal alloy outer shell and an UHMWPE liner. Accordingly, these bearings are metal-on-metal and metal-on-polyethylene. The femoral head can also be ceramic if the acetabular liner is ceramic or polyethylene. Highly crosslinked UHMWPE has been developed to increase polyethylene wear resistance. Fixation betweenthecomponentsandbonecanbeachievedeitherwithimmediatepressfit(frictionfit,interferencefit)followedbyadirectboneingrowthintoporousmetalorhydroxyapatite coating of the implant surface, or with “bone cement”, PMMA, which adheres into the implant surface and interlocks with surrounding bone trabeculae andendocortex.Cementfixedimplantsmayhaveashapecloseddesignandaremeant to be contained by the cement mantle. An alternative is a force closed implant design, relying on subsidence under load as a method of maintaining stability of the cemented implant (Huiskes et al., 1998). A recent review describes the evolution of hip replacement biomaterials from the late 19th century to near future (Wang W et al., 2011). Hip resurfacing is generally not considered to be total hip replacement since in this method only the destructed femoral articular surface is replaced on femoral side. Hip resurfacing replacement and metal-on-metal THR patients were not included in this thesis work. According to recent research, they suffer from early metal debris and ion problems with pseudotumour formation and aseptic loosening (Huo et al., 2011; Matthies et al., 2012; Konttinen and Pajarinen, 2012).

THR complication rate is generally low and most patients experience a marked improvement in quality of life. Long-term survival of the THR above 80% at 20 years has been reported in many series. Comorbid diseases affect the survival of THR (Jämsen et al., 2013). Aseptic loosening of the components has been the most common cause of clinical failure in about 75% of revision operations. Next most common reasons for revision operation are infection (7%), recurrent dislocation (6%), periprosthetic fracture (5%) and technical errors during the primary operation (3%) (Holt et al., 2007). There seems to be a shift in these numbers due to better surgical and cementing techniques and improved implant design and manufacturing. Currently, about 50% of revisions are made because of osteolysis and component loosening, or periprosthetic femoral fractures that are often related to osteolysis and femoral component loosening (Marshall et al., 2008; STAKES, 2013). Swedish and Finnish implant registers show that problems leading to revision operations are often related to certain implant models and technical errors during the primary operation may well be over 3% (SHPR, 2012; STAKES, 2013).

23

aseptic loosening

Osteolysis and aseptic loosening appear to refer to distinct clinical entities. However, they actually represent similar processes with the same basic mechanisms. Osteolysis is in part the result of a particle-induced biologic process at the metal-bone or cement-bone interface, resulting in bone loss. Manifestations of this type of bone loss in patients range from slowly progressing radiolucencies around previously well-fixedimplants,whichmayeventuallyresultinmechanicalinstability,torapidlyexpanding focal lesions that may or may not result in loosening. In orthopaedic vernacular, the former process is generally termed “aseptic loosening” and the latter as “osteolysis”. At the early stage aseptic loosening may be symptomless, but with mechanical instability pain arises. Radiographic appearance shows the development of a ‘lucent line’ around the implant (Figure 2.)

Figure 2. THR with cemented stem and sementless cup 5 years after the implantation (A) and 10 years after the implantation (B). In Figure 1B radiolucent lines are visible especially around the cement mantle. Both components were discovered to be loosened in revision operation.

24

PMMA fragments were described in 1974 in periprosthetic tissues in wear-mediated osteolysis (Willert et al., 1974). Later, particulate PMMA was described around both stable and loose prostheses, in specimens from focal areas of osteolysis and in areas of more linear and uniform bone resorption, giving rise to the concept of “cement disease” (Jones and Hungerford, 1987). In response to this, cementless fixationwasintroduced.Cementlessfixationseemstoworkwell forexample inyoung RA patients (Mäkelä et al., 2011a) but it does not eliminate aseptic loosening and osteolysis (Santavirta et al., 1991; Maloney et al., 1993) and sometimes it is associated with increased aseptic loosening (Eskelinen et al., 2006; Mäkelä et al., 2011b). Subsequently, the term “cement disease” has been replaced by the more accurate “particle disease”. Osteolysis and aseptic loosening is now largely attributed toinflammationinducedbythelocalaccumulationofpolyethyleneandmetalwearparticles and occurs with all materials and in all prosthetic systems that have been used to date (Revell, 2008). Particulate wear debris is especially produced from soft polyethylene, but also metal, bone cement and ceramic materials can produce particles causing foreign body reactions which lead to osteolysis and aseptic loosening (Willert et al., 1990; Buly et al., 1992; Harris, 1995).

The exact causes of aseptic loosening of THR are still somewhat unclear. Surgical technique, implant materials and design clearly affect the risk. There is no clear evidenceof thesuperiorityofeithercementedorcementlessfixation(Pakvisetal., 2011). In the Swedish hip arthroplasty register cemented stems seem more vulnerable to aseptic loosening compared to uncemented stems but cemented cups resist aseptic loosening compared to uncemented cups (Hailer et al., 2010). Aseptic loosening is associated with male sex and young age, possibly due to higher level of physical activity and load. Short stature is a protective factor but risk ratios related to body weight are somewhat controversial. Genetic factors, such as single nucleotidepolymorphism,relatedtoe.g.pro-inflammatorycascadescanaffecttherisk. Perhaps multi-joint OA increases loosening risk this way (Beck et al., 2011). Following factors also contribute to THR aseptic loosening: localized mechanical stresses and cyclic loading (Carlsson et al., 1983; McEvoy et al., 2002); micromotion between the cement or implant and bone (Scott et al., 1985); abrasion particles fromartificialjointsurfaces(Revell,2008);fragmentationofcement(Maloneyetal., 1990); hypersensitivity and other adverse effects (like the lipopolysaccharide-like effect of cobalt Co2+ ions and lysosomal stress and rupture) to metal (Evans et al., 1974; Jacobs et al., 2006; Konttinen and Pajarinen, 2012). Wear particle generationisundoubtedlyasignificantfactorinthestimulationofperiprostheticinflammationandsubsequentboneloss.Itcanberesponsiblealsofor localizedosteolysiseveninwell-fixedprostheses(Wanetal.,1996;Stulbergetal.,2002).

25

Histopathology in aseptic loosening

During function of THRs wear particles are generated as a result of the relative motion and shear between two opposing surfaces under load, as a result of adhesion, abrasion, third body wear, fatigue and corrosion. The following types of particleshavebeenidentified:bonecementdebris,consistingprimarilyofPMMAand x-ray contrast medium (barium sulphate or zirconium oxide), polyethylene particles (conventional or highly cross-linked), zirconia or alumina ceramic wear particles and metal particles. Particles smaller than 150 nm are ingested by endo- orpinocytosis.Particlesrangingfrom150nmto10μmcanbephagocytosedbyosteoblasts,fibroblasts,ECsandinparticularbymacrophages,whilelargerparticlesare surrounded by multinucleated giant cells and foreign body granulomas.

Thecellsareunable todegradewearparticles.Resultingpro-inflammatoryresponse is dependent on chemical reactivity of the particles, on their morphology (fibresaremorepro-inflammatorythanroundparticles),concentrationandsize,with0.24 to7.2μmparticlesbeingmostpro-inflammatory.Metal ion releaseparticularly from metal-on-metal bearings is well documented. It can cause for example metal hypersensitivity, aseptic loosening and formation of pseudotumours (Purdue et al., 2007; Tuan et al., 2008; Hallab et al., 2009; Matthies et al., 2012; Konttinen and Pajarinen, 2012).

Wear debris problem was recognized early by Charnley (1979), when the polytetrafluoroethylene(TeflonR) acetabular components he used required revision within 1 to 3 years in spite of the low friction. After switching to polyethylene cups, Charnley (1968) noticed a dramatic decrease in the amount of wear debris generated and also in the need for revision. In those cases that did require revision, Charnley foundinvariablyfibroustissues,includingmacrophages,surroundingtheimplantat the cement-bone interface (Charnley et al., 1964; 1968). Initially, this interfacial membrane was believed to form in response to the curing of acrylic cement, e.g. toxic effects on non-polymerized methacrylate monomers and damage caused by high temperature due to the exotermic polymerization. However, after evaluating tissues around failed prostheses, Willert and Semlitsch (1976; 1977) proposed that loosening results from the macrophage response to wear debris. Their additional studies suggest that this foreign body-type reaction may contribute to loosening (Willertetal.,1974).Sincethatnumerousclinicalreportshaveverifiedtheirconcept(Jasty et al., 1986; Santavirta et al., 1990; Maloney and Smith, 1995).

A special synovial-like membrane develops at the bone-cement or bone-implant interface in aseptic loosening, possibly as a result of a surface contact of theconnectivetissueimplantcapsulewithsynovialfluiduponexpansionoftheeffective synovial space (Goldring et al., 1983; Lennox et al., 1987; Konttinen et al., 2005). The pathology of the interfacial membrane has been studied extensively. Numerous studies have been reported on:

26

1. Histopathological evaluation of the cellular/stromal makeup. Tissue macrophages,fibroblasts,foreignbodygiantcells,andECsofblood vessels and lymph vessels are frequently observed at the IT while lymphocytes, plasma cells, and polymorphonuclear neutrophilic leukocytes are only occasionally present. The most dominant cell types arefibroblastsandmacrophages(Mirraetal.,1982;Goldringetal., 1983; Lennox et al., 1987; Schmalzried et al., 1992a).

2. Analytic characterization of particulate debris and metal ion burden (Huo et al., 1992; Betts et al., 1992; Cook et al., 1994; Bloebaum et al., 1996; Kadoya et al., 1998).

3. Biochemical analysis of cellular mediators secreted (Goldring et al., 1983; Dorr et al., 1990; Jiranek et al., 1993; Kim et al., 1993; Santavirta et al., 1993; Chiba et al., 1994; Shanbhag et al., 1995; Konttinen et al., 1996b; Hukkanen et al., 1997; Holt et al., 2007).

Goldring et al. (1983) described the histological characteristics of the interface membrane as “synovial-like”. They recognized three distinct zones in the membrane: 1) thesynovial-like liningthat formspapillaryfoldssupportedbyfibrovasculartissue;2)containingalayerinfiltratedbyhistiocytes,giantcellsandparticlesofUHMWPEandPMMA;and3)afibrocartilaginouslayerthatblendsintoadjacentbone.TheyconfirmedtheassociationofhistiocytesandgiantcellswithPMMAparticles at the site of bone lysis.

The joint capsule, which is usually almost completely removed during prosthesis implantation, restores itself within 6 months as a pseudocapsule with a synovial lining, resembling the original. Willert and Semlitsch reported in 1977 their classic observations after removal of failed joint prostheses. They suggested that the whole environmentoftheartificialjointparticipatesintheclearanceofprostheticweardebris. The pseudocapsule appears to be the primary location for phagocytosis of particles. This process results in the development of foreign-body granulomas with areasofnecrosisandfibrosisthataresomewhatproportional totheamountofparticles. Willert and Semlitsch also found that capsular tissue has some capacity to transport particles through the lymphatic system by means of perivascular lymph spaces, leading to regional and systemic distribution of wear debris. Particles accumulate in the periprosthetic tissue if the capacity for elimination by this mechanism is exceeded.

Morphological studies have revealed several similarities between the pseudocapsule and interfacial membrane (Willert et al., 1974; Goldring et al., 1983). In agreement with these observations, both the interfacial membrane and the pseudocapsule have been shown to produce factors that stimulate bone resorption in vitro (Goldring et al., 1983; 1986; Appel et al., 1990; Ohlin et al., 1990;

27

Jiranek et al., 1993). Ohlin and Lerner (1993) have shown that bone resorbing activity is significantlyhigher in thepseudocapsulescomparedwith interfacialmembranes. Consequently, they have suggested that the pseudocapsule may be the most important tissue involved in periprosthetic bone resorption. Pseudocapsule seems to produce osteolytic factors that reach the implant/cement–bone interface insynovialfluidviatheexpandedeffectivesynovialspace.Sub-sequentially,localizedboneresorptionisinitiatedasthefirststepinprostheticloosening.Laterinthisthesis a term “periprosthetic tissue” (PT) is used to cover both pseudocapsule and interfacial membranes.

Schmalzried et al. (1992a) explained that particulate wear debris present in the jointfluidiscarriedtoallperiprostheticregionsthatareaccessibletothefluid.Variationsinthepressureofthefluidandtheflowpatterndeterminewhetherweardebris is linearly dispersed along the component or accumulates in focal areas. The pattern of debris deposition and subsequent biological response would determine whether bone loss is linear or focal. In addition to wear particles, also soluble factors inthejointfluidcandirectlyorindirectlyeffectboneresorption.CytokineandMMPlevelsareelevatedinthefluidaroundTHRsandcouldaffectboneresorptionandtissue destruction at distant sites from the articulating surfaces (Dorr et al., 1990; Takagi, 1996).

Macrophage and fibroblast activation and their role in aseptic loosening

Macrophages are mononuclear cells derived from haematopoietic stem cells along the myeloid lineage. Macrophage survival and proliferation and differentiation of monocytes to macrophages are dependent on M-CSF. The two main cell types in PT are monocyte/macrophages (containing CD68 and CD163 surface marker) and fibroblasts (containing vimentin).Monocyte/macrophages are recruitedfrom the blood stream by the induction of chemokines produced in PTs, while fibroblastsproliferateinsitu(Ishiguroetal.,1997).Theybothcanbeactivatedbywearparticlephagocytosis toproduce inflammatorymediators like tumournecrosisfactor-α(TNF-α),interleukin1β(IL-1β),IL-6,PGE2,M-CSFetc.(Xuetal., 1997; Purdue et al., 2007; Koreny et al., 2006). Those, as well as numerous similar cytokines produced by activated monocyte/macrophages affect not only macrophagesandfibroblasts,butalsoOBsand/orOCsinPTs(seereferencesafewparagraphs above and the literature review of growth factors). A positive feedback cycle develops as the activated cells produce more of the very same chemokines andpro-inflammatorysubstances(Becketal.,2011).ActivatedmacrophagesandfibroblastsassistinOCformationandactivationwithsubsequentboneresorption.Also, activated, debris-related macrophages are capable of direct low-grade bone

28

resorption in vitro (Athanasou et al., 1992). This may be relevant in implant loosening since macrophages seem to cover a larger (19%) surface area than osteoclasts (8%) around of the bone surface surrounding aseptically loosened THRs(Kadoyaetal.,1998).Alsofibroblastsactivatedbytitaniumparticleshavebeen found to produce bone resorbing metalloproteinases (Yao et al., 1995).

In a monocyte culture model where granuloma formation was induced by added particulate material, alkaline phosphatase (AP) activity was found together with cytoplasmic TRAP in CD68+ macrophages (Heinemann et al., 2000). Furthermore, in explants from granulomas obtained from endoprosthetic revisions the major cell type was somewhat surprisingly AP+ cell co-expressing CD68 (Heinemann et al., 2000). AP is a marker for osteoblast and TRAP for osteoclast. In another study, osteoblastic proteins like osteocalcin and osteopontin were produced by macrophagesbutalsobyspindleshapedfibroblast-likecellsderivedfromPTaroundloosenedTHRs(Zreiqatetal.,2003).Thefindingssuggestthattheremaybesomeattempt to transdifferentiation or overlap between tissue macrophage, OC and OB functions in aseptic loosening.

Recently discovered stellate periosteal and endosteal tissue macrophages (osteomacs, see Figure 1.) constitute one sixth of total cells in bone tissues. They are capable of phagocytosis, support OB function, participate in bone remodelling and are involved in bone healing (Pettit et al., 2008; Alexander et al., 2011). They have been suggested to be OC precursors in pathological conditions, to produce pro- and anti-osteoclastic cytokines regulating OC generation and function, and to participate in bone remodelling. Osteomacs may be bone anabolic under physiologic conditionsandcatabolicunder inflammatory/ infectiveconditions(Pettitetal.,2008). Their role in aseptic loosening has not been studied so far.

Osteoclast formation, activation and role in aseptic loosening

After activation, macrophages take part in OC formation. PT-derived chemokines, suchasmacrophage inflammatoryproteinsCLL2,CCL5andCCL3areneededto recruit mononuclear OC precursor cells from blood stream and to initiate their fusion into giant multinucleated cells, foreign body giant cells or OCs. Local macrophages already present in interface tissues are also able to differentiate into OCs (Sabokbar et al., 1997). RANK is part of the TNF receptor family and acts as RANKL receptor on monocyte/macrophages/OC precursor cells. In the presence of M-CSF (Yoshida et al., 1990) it initiates osteoclastogenic signal transduction afterligationwithRANKLexpressedbyOBs,fibroblastsandbonemarrowstromalcells or present in soluble form (Yasuda et al., 1998; Nakagawa et al., 1998). In addition to macrophages, OCs are also capable of wear particle phagocytosis (Wang

29

W et al., 1997). Eventually, activated OCs resorb periprosthetic bone leading to implant loosening.

A combination onM-CSF and RANKL is sufficient for osteoclastogenesis(Quinn et al., 1998). However, even without RANKL, OC precursors are capable ofdifferentiatingintoOCsinthepresenceofM-CSFandTNF-α,TGF-β,IGF-I,IGF-II, IL-6, IL-8 and IL-11 (Kudo et al., 2003; Knowles and Athanasou, 2009; Hemingway et al., 2011). Without M-CSF, OCs can differentiate in the presence of RANKL and HGF, PlGF, VEGF, and Flt3 ligand (Kudo et al., 2003; Knowles and Athanasou,2009;Hemingwayetal.,2011).M-CSF,RANKL,TNF-α,TGF-β,IGF-I,IGF-II, IL-6, IL-8, IL-11 and VEGF are all present in PT tissues (Chiba et al., 1994; Shanbhag et al., 1995; Konttinen et al., 1997a; 1997b; Xu et al., 1997; 1998, Haynes et al., 2001; Jell et al. 2001; Waris et al., 2004).

M-CSFandRANKL,producedbyPTs,areabundantalsoinpseudosynovialfluidwhich penetrates between the implant or PMMA cement mantle and bone perhaps as a result of and during cyclical loading of aseptic loosening THR (Schmalzried et al., 1992b; Xu et al., 1997; Takei et al., 2000; Mandelin et al., 2005a). OBs and ITfibroblastscanassist inOCdifferentiation(Haynesetal.,2001;Sakaietal.,2002; Mandelin et al., 2005b). Systemic and local factors like active vitamin D3, parathyroidhormone,prostaglandinE2,TNF-α,IL-1,IL-6andIL-11assistinOCdifferentiation by inducing RANKL production in OBs and stromal cells also in aseptic loosening (Hallab et al., 2009).

Osteoprotegerin (OPG) is a soluble decoy receptor for RANKL and a physiological negative regulator of osteoclastogenesis by blocking the RANK-RANKL interaction between OC progenitor and RANKL+ cells (Simonet et al., 1997). It is produced by OBs, stromal cells and vascular ECs and inhibits OC formation induced by pseudosynovialfluidfromasepticallyloosenedhipimplants(Itonagaetal.,2000).RANK-RANKL-OPG system is up-regulated in aseptic loosening (Mandelin et al., 2003)andITfibroblastsarecapableofproducingbothRANKLandOPG(Mandelinet al., 2005b).

Osteoblasts in aseptic loosening

The role of osteoblasts (OBs) in aseptic loosening has not been studied as well as that of OCs. Osteoclastic bone resorption is always coupled to osteoblastic bone formation in vivo. Particulate wear debris can induce human mesenchymal stem cell (hMSC) apoptosis and disrupt the differentiation process of hMSC into OBs (Wang M et al., 2004; Tuan et al., 2008). Debris can also be phagocytosed by OBs, induce their apoptosis and reduce their proliferation and bone forming activity (Purdue et al., 2006). However, OBs are present in high numbers in aseptically loosened

30

THR IT samples with adjacent bone where they covered 33% of total bone surface as OCs covered only 8% and macrophages 19% (Kadoya et al., 1998).

Studies on periprosthetic bone around loose prostheses have shown an accelerated turnover with abundant osteoclastic surface but also signs of increased bone formation. This results in immature and bad quality bone near monocyte/macrophage cell sheets in the granulomatous tissues. High bone turnover together with cyclic loading of the implant interface probably contributes to weakening and lossofosseointegration(Takagietal.,2001).SynovialfluidfromasepticallyloosenedTHRs has been found to inhibit the proliferation of OBs (Andersson et al., 2000) but can stimulate OB function (Tsai et al., 2009). The overall effect of cytokines insynovialfluidregulatingboneOBsinsynovialfluidmaychangeaccordingtotheir relative concentration (Andersson et al., 2000). OBs are able to produce bone resorption stimulating cytokines (Wang CT et al., 2010). Furthermore, expression of RANKL by OBs is increased and the OPG/RANKL ratio is decreased in the synovial fluidinasepticlooseningcomparedtocontrolsamples,suggestinganimportantrole for OBs in OC activation in aseptic loosening (Wang CT et al., 2010).

Angiogenesis in aseptic loosening

Angiogenesis is important in bone formation, wound healing and in tissue regeneration and maintenance. It is a complex process where new capillaries form from the pre-existing vasculature. First, in response to angiogenic stimuli, ECs become activated and penetrate through the basement membrane around an existing vessel and invade the surrounding ECM. Then the ECs proliferate and migrate towards chemotactic and angiogenic stimuli and form sprouts. Next, the ECs form tubular structures with a lumen and at the end the new capillary stabilizes byrecruitmentofpericytesanddepositionofbasementmembrane.bFGF,TGF-βand VEGF are involved in the regulation of angiogenesis (Ucuzian et al., 2010).

Stimulated angiogenesis provides nutrition for the developing and remodelling PTs, which suffer from loss of branches of arteries of the femoral head and neck and of the intramedullary femoral capillaries as a result of the surgical THR procedure. Circulating monocytes and OC precursors are able to reach the PT and the periprosthetic bone mainly through post-capillary venules. The vascular ECs take actively part in these processes as well as in the differentiation and activation oftheITfibroblasts,monocyte/macrophagesandOCs(seeabove).Highdegreeofvascularisation was described in the PTs by Jell and Al-Saffar (2001). However, the authors reported poor vascularisation in areas with high levels of wear debris and inflammation.Accordingly,ourgrouphasfoundirregular,reducedanddamagedmicrovasculature in the PTs, possibly in part via ischemia-reperfusion (Santavirta

31

etal.,1996).Itseemslikelythatweardebrisandforeignbodyinflammationdisturbthe angiogenesis in PT, which is characteristic for periprosthetic osteolysis (Collin-Osdoby et al., 2002; Ren W et al., 2006; 2007; Tunyogi-Csapo et al., 2007).

Table 1. Summary of the effects of osteogenic growth factors on bone formation/resorption, angiogenesis and wound/tendon/ligament healing.

effect on bone formation effect on bone resorption

effect on angiogenesis effect on wound / tendon / ligament healing

bFgF

OB proliferation OB differentiation survival of immature OBs chemotactic for OBsbone mineralization

induces osteocyte apoptosisOB differentiation OC formation OC recruitment OC activation

EC migration EC proliferationtube formation ECM remodelation ECM protein synthesis

progenitor cell recruitment migration of fibroblasts differentiation of fibroblasts fibroblast proliferation collagen synthesis

tgF-β

OC proliferation (high concentration)OC differentiation and activation chemotactic for osteoprogenitor cells OB proliferation OB survival ECM protein formation

OC differentiation OC proliferation (low concentration)

inflammatory cell recruitment EC proliferationEC migration tube formation vessel stabilization

inflammatory cell recruitment collagen synthesis tissue debridement by macrophagesECM protein formation

pdgF

OB progenitor migration OB proliferation OB differentiation chemotactic for OBsOB survival

bone resorption OC proliferation OC differentiation

pericyte recruitment pericyte proliferation EC proliferation vessel stabilization

chemotactic for mesenchymal stem cellsmesenchymal stem cell proliferation tissue debridement by macrophages ECM protein formation

igF-i

OB proliferation bone matrix synthesis OB survival

OC differentiation bone resorption

EC migration neovascularization

fibroblast migrationECM protein formation

igF-ii

OB proliferation OB differentiation OB survival

OC differentiation OC activation

EC migration neovascularization

ECM protein formation

Bmp-2,-4,-6,-7

chemotactic for OBsOB proliferation OB differentiation bone matrix synthesis

OC survival OC activation OC function

vascular homeostasis vascular remodelation angiogenesis

chemotactic for mesenchymal cellscollagen synthesis keratinocyte differentiation (BMP-6)

VegF

chemotactic for OBsOB proliferation OB differentiation OB survival OB function endochondral ossification

chemotatic for OC precursorsOC precursor proliferation chemotactic for OCsOC differentiation bone resorption

chemotactic for ECsEC proliferationEC activation EC survival tube formation acts as a key regulator

coordinates ECM degradationacts via angiogenesis regulation

32

Bone Formation enhancing growth Factors

BasicFGF,TGF-β1and-β2,BMP-2,-4,-6,and-7,PDGF,IGFIandIIandVEGFwere selected in this thesis work. Their effects on bone formation and resorption, angiogenesis and wound, ligament and tendon healing are presented in Table 1. They all have osteogenetic potential and thus probably play a role in osseointegration of THRs. Much more attention in aseptic loosening has been paid to osteolysis although it is known that the net loss or gain of bone is dependent on the balance between osteolysis and bone formation. In the simplest scenario it might be so that cytokines and growth factors able to stimulate bone formation are decreased compared to control samples, suggesting a failure of bone formation as an important contributor to peri-implant bone loss. However, the real situation is like to be much more complicated because all these cytokines have also other important effects on bone and on peri-implant tissues. Second, if the osteolysis indeed would be the primary force driving aseptic loosening, it is possible that due to the coupling of OC and OB function, also bone formation would be secondarily (reactively) activated.

Interface membrane development can be regarded as a special type of “internal” wound healing or tissue regeneration process and it is closely associated with periprosthetic osteolysis. Wound healing and tissue regeneration can be divided into1)aninitialinflammatoryphase(protectionandcleaningoftissues)followedby 2) a cellular and vascular proliferative phase and synthesis and deposition of ECM(fillingofthevoidintissues)and3)aremodellingphasewithrestructuringof the ECM and restoration of the tissue function (maturation of the scar or true regeneration of the wounded tissue). The complex sequential steps in this process include migration of cells into the wound / damaged tissue site, proliferation of different cell types, maturation and secretory activities of the cells in which process both the resident and immigrant cells participate (Wang JS, 1996).

Basic FiBroBlast growth Factor

Fibroblast growth factor (FGF)was initially identified andpurified 1973 frompituitaryglandasaprotein thatwasmitogenic forfibroblasts (Armelin, 1973;Gospodarowicz, 1974). Since then, altogether 23 different human FGF subtypes havebeenidentified(Barrientosetal.,2008).ThefirstdiscoveredandmoststudiedFGFs are acidic FGF (aFGF or FGF-1) and basic FGF (bFGF or FGF-2). Basic FGF isproduced inmosthuman tissues.CellsproducingbFGF includefibroblasts,monocytes, macrophages, bone marrow stromal cells, chondrocytes, osteoblasts and endothelial cells (Cook et al., 1990; Devescovi et al., 2008). Usually FGFs bind to cell surface (e.g. syndecan and glypican) or ECM (perlecan and agrin) heparan

33

sulphate proteoglycans, which localise and protect them so that they can be released and activated from these storage sites upon tissue degradation (Barrientos et al., 2008; Yun et al., 2010). Basic FGF absorbed to ECM can be released by various enzymes, like heparinases (Wang JS, 1996). FGFs in a complex with heparin and heparin-like poly- or oligosaccharides act by binding to target cell tyrosine kinase FGF receptors, four of which have been described in humans (Barrientos et al., 2008; Yun et al., 2010). FGFs have multiple functions in embryogenesis and tissue regeneration affecting target cell proliferation, migration and differentiation. Target cells are, depending on the FGF subtype, endothelial, epithelial and neural cells, fibroblasts,OBs,osteocytes,OCs,keratinocytesanddifferenttypesofstemcells.Basic FGF enhances the regeneration of bone, vessels, tendon, and cartilage as well as neural and periodontal tissues in both in vitro and in vivo animal models. Human experiments have not been documented so far. Free FGFs have a short half-life and thus bFGF is usually delivered within different types of matrices in animal experiments (Yun et al., 2010).

Effects on bone

Acidic FGF, bFGF, FGF-8, FGF-9 and FGF-18 are all osteogenic, but bFGF is by far the most studied of them in skeletal development, bone physiology and fracture healing (Shimoaka et al., 2002; Kelpke et al., 2004; Lin et al., 2009; Yun et al., 2010; Behr et al., 2010). Osteoblast derived bFGF is stored in particular in bone matrix and can be released from there. Basic FGF assists in recruitment and proliferation of OB progenitor cells and affects bone remodelling by inducing OB proliferation and differentiation and promoting their survival (Hill et al., 1997; Collin-Osdoby et al., 2002; Park, 2011). It has been shown to exhibit in vitro chemotactic activities towards primary osteoblasts (Mayr-Wohlfart et al., 2002). Depending on the circumstances, it is able to either stimulate or inhibit OB differentiation (Collin-Osdoby et al., 2002). Basic FGF has been found to have positive effects on bone mineralization and regeneration, fracture healing and implant osseointegration both in vitro and in vivo in numerous animal models (Wang JY 1996a; Yun et al., 2010; Kaigler et al., 2011).

Also clear bone formation inhibiting effects have been found. Namely, bFGF appears to inhibit the differentiated function of OBs for example by suppressing IGF-I synthesis (Canalis, 2009). Furthermore, bFGF stimulates OC recruitment and is able to induce osteoclastogenesis and OC activation via RANKL leading to resorption pit formation (Collin-Osdoby et al., 2002; Shimoaka et al., 2002).

34

Effects on angiogenesis

Angiogenesishasanimportantroleininflammationandtissuerepairandinboneformation. In addition to bone matrix, bFGF is also stored in the ECM of the vascular basement membrane. It promotes early phase of angiogenesis by mediating proteolysis of ECM components enabling EC migration. Later, it induces endothelial cell proliferation and the organization of endothelial cells into tubular structures. Furthermore, it affects ECM remodelling during angiogenesis by inducing ECs to synthesizecollagen,fibronectinandproteoglycans.BasicFGFbindstoECMproteins,which further potentiate its effects locally upon progress of the angiogenesis. It is a stronger angiogenic factor than VEGF or PDGF, but they all act synergistically modulating effects of each other on angiogenesis (Ucuzian et al., 2010; Yun et al., 2010). Basic FGF has positive effects on angiogenesis in vitro and in vivo in animal models (Yun et al., 2010).

Effects on wound, ligament and tendon healing

Basic FGF is also heavily involved in wound, ligament and tendon healing with its expression peaking approximately one week after injury (Isaac et al., 2012). It is released by endothelial cells and macrophages at wound/injury sites (Yun et al., 2010). If bFGF activity is blocked, wound angiogenesis fails almost completely. It has proangiogenic effects in wound healing, induces recruitment of progenitor cellsandmigrationanddifferentiationofmultiplecelltypeslikefibroblasts.Italsoinducesfibroblastproliferationandcollagensynthesisrequiredinwoundhealingandfibroustissueregeneration(Molloyetal.,2003;Yunetal.,2010).BasicFGFisproducedbyinflammatorycellsandfibroblastsnecessaryinwoundhealingandconnective tissue development (Chang et al., 1998; Cool et al., 2004; Barrientos et al., 2008; Longo et al., 2011). Basic FGF has been found to have positive effects on wound and tendon/ligament healing in vitro and in vivo in animal models (Yun et al., 2010; Isaac et al., 2011; Longo et al., 2011; Pastides and Khan, 2012).

transForming growth Factor-βs

The three known TGF-β isoforms TGF-β1, TGF-β2 and TGF-β3 are secretedproteinsthatbelongtotheTGF-βsuperfamily.This isa largefamilyofgrowthand differentiation factors with over 40 members, including bone morphogenetic proteins (Chenet al., 2012).TGF-βsare synthesizedand secretedbymultiplecell types like fibroblasts, endothelial cells, keratinocytes, platelets, OBs andmacrophages(Barrientosetal.,2008).TGF-βisproducedasaninactiveprecursor

35

molecule containing a propeptide region. It is homodimerised and after furin-mediatedcleavageofthepropeptide(whichremainsboundtoTGF-β)complexedwith a latency associated peptide. After further complex formation with a latent TGF-β-bindingprotein,itissecretedasalargelatentcomplex.ThisprecursorformofTGF-βisstoredinECM.ThelargestreservoirsofTGF-βsareboneandplatelets(Allorietal.,2008).TGF-βsandIGF-IIarethemostabundantgrowthfactorspresentinhumanbone(Mohanetal.,1990).TGF-βcanbeactivated(released)byserumproteinases, low pH and reactive oxygen species (Allori et al., 2008). Active and free TGF-βdimerstransmitsignalsacrosstheplasmamembranethroughbindingtotypeII receptors, which homodimerise, and then form heterotetrameric complexes with typeI(ALK5)receptordimers.BothtypeIandIITGF-β-Rareserine/threoninekinases. Activated receptors initiate intracellular signalling through phosphorylation of receptor regulated SMAD proteins, which form a complex with a co-SMAD, e.g. SMAD4. Such complexes then translocate into the nucleus to direct transcriptional responses(Chenetal.,2012).TGF-βsplayrole forexample inembryogenesis,connective tissue development and healing, vascular development, immunity, canceranddiabetes.TGF-βsarepleiotropiccytokineswithwell-knownpro-andanti-inflammatoryproperties(Sanjabietal.,2009).

Effects on bone

Depending on local and systemic circumstances, TGF-βsmay be anabolic orcatabolicforbone.AllTGF-βisoformsappearcapableofinducethesameeffectson bone but their potency varies according to receptor binding capacity (Allori et al.,2008).OB-derivedTGF-β1and-β2arestoredintheECMasa latent formin large latent complexes linked to proteoglycans and glycoproteins during the formationofECM.BoneresorptionbyOCscanreleaseactiveTGF-βtoassistforexampleinfracturehealing.InadditiontothestoredandreleasedTGF-β, it isalso highly expressed by many cell types during new bone formation in fracture healing (Allori et al., 2008).

TheeffectsofTGF-βchangeduringthedifferentphasesofboneformation.Initially, it is chemotactic for osteoprogenitor cells. At the same time it inhibits the proliferation, differentiation and activation of OC precursors and OCs (Allori et al., 2008). Later it increases OB proliferation and survival while being a potent promoter of collagen production and early differentiation of osteoblast-like cells (Bonewald, 1996; Jilka et al., 1998; Ogino et al., 2006; Allori et al., 2008). In a later phase, it is involved in the regulation of mineralization via up-regulation of non-collagenous ECM proteins (Allori et al., 2008). During bone remodelling phaseTGF-βactsasapromoterofosteoclastogenesisandOCactivationwhereby

36

more ECM-boundTGF-β is released (Allori et al., 2008). It is also a potentchemoattractant for OCs (Pilkington et al., 2001).

Insomecircumstances,TGF-β1inhibitsboneresorption(Takaietal.,1998)andalsoTGF-β2hasbeenfoundtodecreaseRANKLlevelsandtoincreaseOPGlevels,thus interrupting the main osteolytic pathomechanism (Holt et al., 2007, Mandelin etal.,2005b).Ontheotherhand,over-expressionofTGF-β2hasbeenfoundtocontribute to bone loss and to associate with defective bone mineralization and remodelling in mice (Erlebacher et al., 1996).

TheroleofTGF-βinOCformationiscomplexandinsufficientlyunderstood.It may increase OC formation via action on OC precursors. Since it has both positive and negative effects on OC differentiation, it probably acts via more than onemechanism.TGF-βisapowerfulmodifierofosteoclastogenicstimulationbyRANKLandTNF-α,whichareessential forOC formationandactivation(Quinn et al., 2001; Lorenzo et al., 2008; Knowles and Athanasou, 2009). An exampleofthecomplexrolesofTGF-β1inosteolysisisthefindingthatitinhibitsOC formation in rat bone marrow cultures, but those OCs that form become activatedbyTGF-β1andshowincreasedboneresorbingactivity(Hattersleyetal.,1991).However,mostinvivoanimalstudieswithTGF-βshaveregisteredenhancement of bone formation for example in fracture healing, distraction osteogenesis and around metal implants (Allori et al., 2008; Lamberg et al., 2009; Kaigler et al., 2011).

Effects on angiogenesis

TGF-β is able to affect angiogenesis indirectly by recruiting proangiogenicinflammatorycellstothesiteofcapillaryformationandbymodulatingangiogenesispathways (Ucuzian et al., 2010). It also regulates directly EC proliferation, migration and capillary tube formation (Li et al., 2003). Migration of smooth muscle cells and pericytes surrounding newly formed capillaries and EC-pericyte interactions invesselstabilizationphasearealsoregulatedbyTGF-βtogetherwithPDGFandVEGF (Ucuzian et al., 2010).

Effects on wound, ligament and tendon healing

Duringwoundhealing,TGF-β1isimportantinearlyinflammation,angiogenesis,connective tissue regeneration and re-epithelialisation. It assists in recruiting inflammatorycellsandlaterintissuedebridementbymacrophages.Subsequently,it enhances ECM formation and up-regulates VEGF. Later, it is involved in

37

collagenproduction,fibronectinbindinginteractionsandiscapableto inhibitcollagen breakdown by MMPs (Molloy et al., 2003; Barrientos et al., 2008).

TGF-β1and-β2havebeenfoundtohavepositiveeffectsintendon/ligamenthealing in vitro and in vivo in animal models (Longo et al., 2011; Isaac et al., 2012; PastidesandKhan,2012;Yilgor et al.,2012).TGF-β1has improvedmeniscustissue engineering (Yang, 2009) but is over expressed also in harmful processes like tendinopathy and hypertrophic scarring (Longo et al., 2011).

platelet deriVed growth Factor

PDGF was originally identified as a disulfide-linked dimer of two differentglycoprotein chains, A and B, which could be separated using reversed phase chromatography (Johnsson et al., 1982). Since then, various homo- or heterodimeric PDGF isoforms have been found, namely AA, AB, BB, CC, and DD. They belong to the same growth factor family with VEGF (Fredriksson et al., 2004). These isoforms bind withdifferentaffinitiestoandsignalthroughtwodistincthomo-orheterodimerisingreceptortyrosinekinases(αandβ)expressedonthesurfacesofvariouscelltypes(Caplan and Correa, 2011). PDGF-BB is thought to be the universal PDGF because of its ability to bind to all known receptor isotypes (Hollinger et al., 2008) and due to its physiological functions (Caplan and Correa, 2011). PDGFR effects are mediated mainly via phosphoinositide-3-kinase–protein kinase B/Akt pathway (Hemmings and Restuccia, 2012) and Ras/Raf - MEK1/2 - ERK1/2 -pathway (Avruch et al., 2001), but also via phospholipase C gamma (Morrison et al., 1990).

During embryogenesis, PDGFs assist in the formation of multiple tissues/organs like hair, alveoli, glomeruli, cardiovascular system and skeleton by inducing cell replication and directing cell migration. Their role in normal physiology remains unclear, but they are heavily involved in various pathologic conditions like cancer, vasculardiseasesandfibroses(Andraeetal.,2008).PDGF-BBseemstohaveacentral role in bone formation and regeneration because it regulates pericyte, MSC and ECM functions in angiogenesis. It is able to couple angiogenesis and bone formation for example via VEGF and BMPs (Caplan and Correa, 2011). The main source of PDGF are platelets in blood stream, which synthesize, store and burst release it in a regulated manner. However, PDGF is produced also by cells like activated macrophages. Bone and ECM of other tissues contain some protein-bound PDGF. PDGF, as well as many other growth factors, can be released from platelets after their aggregation for example in a fracture site, where it rapidly increases OB proliferation (Canalis, 2009).

38

Effects on bone

PDGF is found in a variety of tissues, including bone, where it acts as a regulator of skeletal remodelling. PDGFs are potent bone cell mitogens that stimulate proliferation of osteoblastic cells similarly to bFGF (Canalis, 2009). PDGF isoforms are powerful chemoattractants for MSCs and OBs (Fiedler et al., 2004; Allori et al., 2008). In vitro, PDGF stimulates OB and OC precursor proliferation but appears to inhibit bone matrix apposition (Hock et al., 1994). It can potentiate the effect of IGF-I and -II on OB survival (Hill et al., 1997).

On the other hand, PDGF-BB has been found to modulate OC differentiation and to promote adult osteoclastic bone resorption directly through PDGF receptor beta (Zhang et al., 1998). PDGF-BB promotes adult osteoclastic bone resorption directly through PDGF receptor beta and plays important roles in bone healing and reconstruction (Zhang et al., 1998). In addition, PDGF-BB probably plays a role in coupling of bone resorption and formation, since OC-derived PDGF-BB has been found to be chemotactic for OBs (Sanchez-Fernandez et al., 2008). PDGF has positive effects on bone formation in a number of in vivo animal studies (Nash et al., 1994; Allori et al., 2008; Kaigler et al., 2011) as well as in prospective human clinical trials to treat periodontal bony defects (Kaigler et al., 2011).

Effects on angiogenesis

PDGF seems to be critical in the recruitment and proliferation of pericytes during angiogenesis.PDGFproducedbyECsischemotacticandmitogenicforfibroblasts,smooth muscle cells and other mesenchymal cells that gather around endothelial cells and stabilize and support them. PDGF is known to induce EC proliferation indirectly for example by up-regulating VEGF expression by ECs and may be one on the key molecules to link angiogenesis and bone formation (Ucuzian et al., 2010;CaplanandCorrea,2011).KnockoutofPDGFhasledtosignificantvascularabnormalities in animal studies (Ucuzian et al., 2010).

Effects on wound, ligament and tendon healing

PDGF plays a role in every phase of wound healing. First, it serves as a chemokine andmitogenforneutrophils,macrophagesandfibroblasts.Subsequently,itassistsin tissue debridement by macrophages, fibroblastic ECM production, tissueremodelling and granulation tissue growth (Barrientos et al., 2008). In tendon healing, PDGF is typically produced in early phase after tendon damage, the expression peaking between 1 and 2 weeks after injury. It acts as a chemokine and

39

mitogen and stimulates the production of other growth factors and has in a later phase a role in tissue remodelling (Molloy et al., 2003; Isaac et al., 2012). It has been found to have positive effects in wound and tendon/ligament healing in vitro and in vivo in animal models (Barrientos et al., 2008; Longo et al., 2011; Isaac et al., 2012; Pastides and Khan, 2012; Yilgor et al., 2012). Recombinant PDGF has been introduced to clinical use to improve human wound healing (Andrae et al., 2008; Barrientos et al., 2008).

insulin-like growth Factors

IGF-I (somatomedin-C) and -II, two highly similar anabolic peptides, are synthesized inmultipletissuesbymanycelltypes,includingconnectivetissuefibroblastsandstromal cells (Cohick and Clemmons, 1993), bone cells (Mohan and Baylink, 1991) andlymphocytes(Geffneretal.,1990).IGF-I,-IIandTGF-βarethemostabundantgrowth factors present in skeletal tissue, the concentration of IGF-II being higher than that of IGF-I (Mohan et al., 1990; Canalis, 2009). Systemic IGF-I is synthesized by liver under the control of growth hormone (GH) but also in local tissues like muscle and bone. It is mostly bound to an IGF-binding complex formed by IGF-I, IGF-binding protein (IGFBP)-3, the predominant circulating binding protein, or IGFBP-5 and the acid labile subunit (ALS). The concentration of IGFBPs is higher than that of IGF-I. Therefore, IGF-I circulates mostly bound to the complex and less than 1% of total serum IGF-I circulates as a free, active hormone (Allori et al., 2008; Giustina et al., 2008; Canalis, 2009). IGFBPs and proteases degrading them regulate the bioavailability of IGFs. Bound IGF-I is inactive and thus the local production of IGF-I is often considered to be more important for regulation of growth at tissue level (Cohick and Clemmons, 1993). Local IGF production is regulated by many trophic hormones and nutrition, but also locally by an array of growth factors and other cytokines. The availability and activity of IGF-I is regulated by six IGFBPs (Canalis, 2009). There are two receptors for IGFs. Both IGF-I and IGF-II bind to the IGF1-receptor, which is a tyrosine kinase, while only IGF-II binds to the non-signalling IGF2-receptor sequestering IGF-II. IGF1-receptor signalling pathways overlap to an extent with those of the PDGF, involving e.g. phosphoinositide-3-kinase and mitogen activated kinase pathways.

IGFs have mitogenic and differentiation promoting effects on multiple cell types,includingfibroblasts,musclecells,osteoblastsandchondroblasts(CohickandClemmons, 1993; Sara and Hall, 1990; Giustina et al., 2008). IGF-I has an important role in the regulation of cell metabolism in many tissues, including cartilage and bone. In addition to their role in normal tissue development, the IGFs play important roles in diseases like diabetes and cancer (Maki, 2010).

40

Effects on bone

IGF-IdoesnotinfluencethedifferentiationofMSCsintoOBsandcellproliferation,which instead is stimulated by IGF-II (Tsiridis et al., 2007). IGF-I enhances bone matrix type I collagen production (Canalis, 2009). It is mitogenic for OBs and stimulates them to form bone at both systemic and local level. IGF-II can also stimulate bone formation but IGF-I is a more potent stimulator at least in vitro (McCarthy et al., 2009). Both IGFs promote OB survival and the effect is potentiated by PDGF and bFGF (Hill et al., 1997). Interestingly, IGF-I seems to play a role in bone resorption since it can induce RANKL and thus osteoclastogenesis, modulate OC differentiation and enhance OC function (Mochizuki et al., 1992). Both IGFs are able to substitute RANKL for OC differentiation and activation in vitro (Hemingway et al., 2011).

IGFs are speculated to couple bone formation and resorption by the following mechanisms (Hayden et al., 1995): 1) OB derived IGFs incorporated into bone matrix are released during osteoclastic bone resorption and cause osteoblastic, site-specificbonereplacement;2)OCderivedIGFs(Mochizukietal.,1992)actonneighbouringOBsandstromalcellsenablingthemtofilltheresorptioncavitieswithbone; 3) OB derived IGFs act as autocrine agents. IGF-I may couple angiogenesis to bone formation since it induces VEGF expression in OBs (Akeno et al., 2002). It has positive effects for bone formation in vitro and in vivo in animal models including peri-implant bone, but also in an early clinical trial for human periodontal regeneration in combination with PDGF (Giustina et al., 2008; Lamberg et al., 2009; Kaigler et al., 2011).

Effects on angiogenesis

IGF-I and -II have been found to be strong enhancers of EC migration and neovascularisation (Su et al., 2003; Grant et al., 1987; 1993). The effects may be partially indirect, since at least IGF-I was found to induce several genes that positively regulate angiogenesis (including VEGF), while repressing many genes that inhibit angiogenesis (Oh et al., 2002).

Effects on wound, ligament and tendon healing

IGF-Iishighlyexpressedduringtheearlyinflammatoryphaseinanimaltendonhealingmodels,andappearstoaidintheproliferationandmigrationoffibroblastsand, together with IGF-II, to subsequently increase proteoglycan and collagen production (Molloy et al., 2003; Longo et al., 2011). As anabolic growth factors,

41

both IGFs have active roles in the healing of different types of wounds (Ching et al., 2011). They have been found to have positive effects in wound and tendon/ligament healing in vitro and/or in vivo in animal models (Ching et al., 2011; Longo et al., 2011; Pastides and Khan, 2012; Yilgor et al., 2012).

Bone morphogenetic proteins

BMPsbelongtotheTGF-βsuperfamilyofgrowthfactorsandareexpressedaslong precursor proteins with an N-terminal signal peptide, a prodomain and a C-terminal mature peptide. Endoproteases cleave the proprotein and produce the mature protein that is subsequently secreted (Cui et al., 1998). Secreted BMPs can bind to cell surface and ECM heparan sulphate proteoglycan storage sites, which can also restrict receptor interactions. BMPs interact with BMP receptors composed of either 1A or 1B in a heterodimeric complex with BMP receptor 2, which receptors are located on the cell surface. Signal transduction through BMP receptors results in mobilization and activation of receptor activated SMADs, which then heterodimerise with common partner (mediator) SMADs to regulate transcription factors in the nucleus.BMPscanalsoactivateMAPKpathwayviaTGF-β1activatedtyrosinekinase1. The signalling pathways involving BMPs, BMPRs and SMADs are important in the development of the heart, central nervous system and cartilage, as well as in post-natalconnectivetissuedevelopment.LocaleffectsofBMPsaremodifiedbyanumberof different antagonists and modulators. BMPs have an important role for example during embryonic development, skeletal formation, bone homeostasis, fracture healing and bone tumour development (Laitinen et al., 1997; Chen et al., 2012). They are synthesized in skeletal tissues by OBs and OCs, but also in many extraskeletal tissuesbyothercells,includingfibroblasts,macrophagesandECs.Theyactasgrowthfactors, morphogens or pleiotropic cytokines depending on their spatio-temporal expression and target cells. In addition, they can affect cell motility and apoptosis. BMPs drive differentiation of MSCs to bone cells and subsequent bone formation. BMP-2, -4, -5, -6, -7 (OP-1), -8, -9 and -10 have osteogenic properties (Lissenberg-Thunnissen et al., 2011). The same BMPs seem to inhibit OC differentiation and activation. BMPs are often regulated by local feedback mechanisms maintaining tissue homeostasis (Rosen and Wozney, 2002; Canalis et al., 2003; Lories et al., 2005; Garimella et al., 2008).

42

Effects on bone

BMPs are the most osteoinductive of all growth factors with BMP-2, -4, -6, -7 and -9 being most osteogenic isoforms. At low concentrations they are chemotactic and mitogenic for MSCs and OBs (Mayr-Wohlfart et al., 2002; Fiedler et al., 2004; Allori et al., 2008). At high concentrations they are capable of inducing osteoblastic differentiationinseveralcelltypeslikeMSCsandfibroblasts.Importantfunctionsof BMPs are the induction of OCs and inhibition of OB maturation (Allori et al., 2008; Canalis, 2009). These functions are needed in bone remodelling which is partly regulated by BMPs. They induce osteoclastogenesis by increasing RANKL production by OBs, promote OC survival and have been found to stimulate OC activation and function (Mishina et al., 2004; Canalis, 2009). However, they are also capable of inducing OPG production by OBs. OPG binds RANKL and thus reduces osteoclastogenesis (Canalis, 2009).

Finnish researchers, together with Marshall R. Urist, “the father of BMP” (Urist, 1965), have reported enhanced bone healing with BMP in the 1980's (Lindholm et al.,1988). Since then, BMP-2, -4, -6 and -7 have shown positive effects on bone formation, fracture healing and implant osseointegration in vitro and in vivo in numerous animal models (Lissenberg-Thunnissen et al., 2011; Wang W et al., 2011; Thorey et al., 2011). BMP-2 improves the healing of human open tibial fractures treatedwithunreamedintramedullarynailfixation(Govenderetal.,2002)buttheeffect disappears with reaming (Aro et al., 2011). BMP-2 and -7 are currently in clinical use to enhance bone formation in delayed bone repair and in spinal fusion surgery where BMP performs better than cancellous bone autografts (Agarwal et al., 2009; Lissenberg-Thunnissen et al., 2011).

Effects on angiogenesis

BMPs and their signalling systems play important roles in the development the embryonic vasculature, but also in adult vascular homeostasis, tissue remodelling, injury and various diseases like general and pulmonary hypertension and atherosclerosis (Lowery et al., 2010). At least BMP-2, -4, -6 and -7 have been found to be clearly angiogenic in vitro (Shao et al., 2009; Ren et al., 2007; Ramoshebi et al., 2000). BMP- 2, -4 and -6 have been found to stimulate the expression of VEGF by OB-like cells (Deckers et al., 2002) suggesting that BMPs and VEGF play a role in coupling of angiogenesis and bone formation (Beamer et al., 2010).

43

Effects on wound, ligament and tendon healing

BMP-2, -4, -6 and -7 are all expressed in wound tissue where BMP-6 has a role in keratinocyte differentiation, but their other functions in wound healing are still unclear (Barrientos et al., 2008). BMP-2 and -7 increase the collagen production of tenocyte-like cells in vitro (Pauly et al., 2012). BMP-2, -6 and -7 have positive effects in tendon/ligament healing in vitro and/or in vivo in a few animal models (Bobacz et al., 2006; Pastides and Khan, 2012). In addition, they have been found to improve osseointegration between ACL and bone tunnel in at least three studies (Yilgor et al., 2012) and, together with BMP-7, it improved tendon/bone junction strength in a sheep infraspinatus repair model (Rodeo, 2007). Furthermore, BMP-4 has showed promising results in meniscus tissue engineering (Yang et al., 2009).

Vascular endothelial growth Factor

Vascular endothelial growth factor family consists of seven secreted members, VEGF-A (also referred as VEGF), VEGF-B, -C, -D, -E and -F and placenta growth factor. Of these growth factors, VEGF-A is the most important angiogenic cytokine and is also able to enhance bone formation (Dai and Rabie, 2007). It is produced bymacrophages,fibroblasts,ECsandothercelltypes(Ferraraetal.,2003,Berseet al., 1999, Seghezzi et al., 1998). VEGF binds to VEGF receptor-1 (VEGFR1/Flt-1) and receptor -2 (VEGFR2/KDR/Flk-1). These receptors are mainly produced byendotheliumbutalsobymonocyte/macrophages,fibroblasts,myofibroblasts,synovial lining cells, OCs and OBs (Neufeld et al., 1999; Ikeda et al., 2000; Sawano et al., 2001; Giatromanolaki et al., 2001; Sato et al., 2003; Chintalgattu et al., 2003; Itoh et al., 2006; Dai and Rabie 2007). VEGF can also bind to neuropilin-2 and to heparan sulfate proteoglycans thus inducing angiogenesis (Dai and Rabie, 2007). VEGF is important during embryogenesis as well as in the development and repair of normal tissue including skeletal tissue. It also has an important role in cancer, diabeticretinopathyandinflammatorydiseaseslikerheumatoidarthritis(Parikhand Ellis, 2004; Dai and Rabie, 2007). The expression of VEGF and its receptors can be regulated by hypoxia, hormones like thyroid hormone, MMP-9, BMP-2, -4, -6and-7,TGF-β1,PDGFs,IGF-IandothercytokineslikeIL-1β,IL-6,IL-10,IL-11and TNF-α (Neufeld et al., 1999; Parikh and Ellis, 2004; Dai and Rabie, 2007).

Effects on bone

VEGF/VEGFR system plays an important role in the regulation of bone biology. Itisinvolvedinintramembranousandendochondralossification,fracturehealing

44

and bone metastasis (Street et al., 2002; Aldridge et al., 2005; Dai and Rabie 2007; Allori et al., 2008). It affects bone remodelling by attracting ECs and OBs and by stimulating OB differentiation and inhibiting OB apoptosis (Mayr-Wohlfart et al., 2002; Deckers et al., 2000; Street and Lenehan, 2009). VEGF can stimulate OB activity directly and independent of a prevailing vasculature (Street and Lenehan, 2009). Experimental studies suggest that VEGF may couple angiogenesis and bone formation by manipulating the angiogenic response to osteoblastic activity (Beamer et al., 2010). Furthermore, VEGF may act as an autocrine regulator of osteoblastic differentiation and activity. By expressing VEGF, OBs may induce nearby cells to express factors that regulate osteoblastic activity (Dai and Rabie, 2007; Zelzer and Olsen, 2005). According to a knockout mouse study, VEGFR1 signalling seems to be important in OB activity (Niida et al., 2005). VEGF has been found to have positive effects for bone formation in vitro and in vivo in many animal studies (Allori et al., 2008; Beamer et al., 2010).

Interestingly,VEGFhasalsosignificanteffectsonosteoclasticboneresorptionand stimulates chemotaxis and cell proliferation in OC precursor cells (Nakagawa et al., 2000; Niida et al., 1999). Monocyte and OC precursor cells chemotaxis is mediated through VEGFR1 and -2. VEGF is known to act on OC differentiation via VEGFR1, whereas VEGF-induced osteoclastic bone resorption is mediated by VEGFR2 (Knowles and Athanasou, 2009; Aldridge et al., 2005; Niida et al., 2005).

To sum up, VEGF can mediate either bone formation or resorption, the ultimate balance depending on cell receptor expression, differentiation state, and the cytokine, biophysical and biochemical milieu.

Effects on angiogenesis

VEGF is a powerful angiogenic growth factor that has numerous inductive effects on proliferation, permeability, invasion, migration, survival and activation of ECs and acts as a key regulator of physiological and pathological angiogenesis (Parikh et al., 2004). VEGF, bFGF and PDGF act synergistically modulating effects of each other in angiogenesis (Seghezzi et al., 1998; Yun et al., 2010; Ucuzian et al., 2010).

VEGFR1 and VEGFR2 are predominantly located on the surfaces of endothelial cells (Neufeld et al., 1999). VEGFR2 seems to be the major mediator of the mitogenic, angiogenic and permeability-enhancing effects of VEGF. VEGFR1 can act as a decoy receptor preventing VEGF binding to VEGFR2. It can also modulate the effect of VEGF on endothelial cell invasion and ECM degradation (Ferrara et al., 2003).

The effects of VEGF on EC invasion as well as its ability to coordinate ECM degradation result from VEGF-based induction of MMPs, urokinase-type plasminogen activator (uPA), urokinase-type plasminogen activator receptor and

45

tissue-type plasminogen activator (Parikh et al., 2004). VEGF induces EC migration, mitogenesis, sprouting and tube formation. It increases vascular fenestration and permeability and acts synergistically with many other angiogenic mediators in different phases of angiogenesis (Ucuzian et al., 2010). VEGF can affect EC migration by activating focal adhesion kinase and the p38 pathway (Parikh et al., 2004). VEGF-induced NO synthesis contributes both to EC migration and permeability (Neufeld et al., 1999; Parikh et al., 2004).

VEGF seems to have an important role in modulating angiogenic effects of RANKL under physiological and pathological conditions mainly through VEGFR2 pathway (Min et al., 2003). Similarly to bFGF, binding of VEGF to ECM proteins enhances its effects (Steffens et al., 2004). As with many other angiogenic mediators, complex interactions between VEGF, ECs, pericytes and ECM proteins are crucial for proper vessel formation (Ucuzian et al., 2010).

Neovascularisation is essential in tumour growth, wound healing, tissue repair andremodellingandplaysanimportantroleinacuteandchronicinflammatoryconditions. Abundant research has been and is being done with VEGF and its receptors in pro- and especially in antiangiogenic therapies. VEGF antibodies are in clinical use in the treatment of cancer and age-related macular degeneration (Goel et al., 2011).

Effects on wound, ligament and tendon healing

Angiogenic properties of VEGF make it crucial in acute and chronic wound healing.Inthesesettings,itisderivedfromplatelets,neutrophils,fibroblastsandmacrophagesandisup-regulatedbyhypoxiaandbycytokineslikebFGF,TGF-β1and PDGF (Barrientos et al., 2008). Similarly, VEGF is a powerful stimulator of angiogenesisintendonhealingaftertheearlyinflammatoryphase(Molloyetal.,2003; Yilgor et al., 2012). Peak levels of VEGF appear at 7-10 days after tendon trauma with majority of cells expressing it at the repair site (Longo et al., 2011). VEGF has positive effects in wound and tendon/ligament healing in vitro and in vivo in animal models and in human ischemic skin ulcer patients (Barrientos et al., 2008; Longo et al., 2011).

46

3 AIMS OF THE STUDY

Current work regarding THR is focused mainly on the cytokines and growth factors, which stimulate bone formation and the failure of which might contribute to a net loss of bone in aseptic loosening. It has been assumed that an implant capsule and later a layer or membrane of synovial-like tissue develops at the implant-to-bone interface almost invariably in aseptic implant loosening and, with time, leads to loosening of the components from the surrounding bone. Researchers have found that this membrane produces various mediators, especially cytokines, which contribute to OC formation and bone resorption in aseptic loosening. During the course of this study the importance of RANK/RANKL/OPG -system has been emphasizedinthefieldofasepticlooseningresearchwhereascytokinesstimulatingOBs and bone formation have received less attention.

A bulk of knowledge has accumulated on cytokines acting on bone tissue. It is now clear that these cytokines act mainly as autocrine or paracrine factors on proliferation or differentiation of bone cells. They are involved in the physiologic and pathological coupling of resorption and formation. When we planned this study our research group and others had already found that the expression of osteolytic cytokineslikeTNF-α,M-CSF,GM-CSF,IL-8,IL-11andTGF-αisup-regulatedintheaseptic loosening. Because less attention has been paid on the expression of bone formation enhancing factors in aseptic loosening, we wanted to focus on cytokines thatareabletostimulateboneformation.Atthesametimeintheresearchfieldof another bone destruction type, osteoporosis research, it was found that skeletal contentofTGF-β1andIGF-Iisdiminishedwithincreasingage(Nicolasetal.,1994).This led to our hypothesis that bone formation enhancing cytokines are decreased in ITs in patients with loose THR compared to control synovial membranes.

Outlines of the study

I. To study the eventual presence, cellular localization and extent of expression of bFGF,TGF-β1and-β2,BMP-2,-4,-6,and-7,PDGF,IGF-Iand-IIandVEGFinthe interface and pseudocapsular membranes retrieved from loose hip prostheses using immunohistochemistry and image analysis.

II. To analyse the presence of IGF-I and -II mRNA in IT using reverse transcription polymerase chain reaction (RT-PCR).

47

III. To study BMP-production by mesenchymal stromal/stem cells and osteoblasts using immunostaining.

IV. To study the regulation of VEGF/VEGR receptor system by hypoxia and cytokines inculturedfibroblastsusingquantitativereal-timereversetranscriptionpolymerasechain reaction (qRT-PCR).

48

4 MATERIALS AND METHODS

patients and samples

Study design was approved by the ethical committee of Helsinki University Central Hospital according to the Declaration of Helsinki. Multiple samples from each patient were collected and then embedded and frozen in OCT compound (Lab-Tek Products) and stored at –20°C until used in immunohistochemistry (studies I-IV) orinRT-PCR(studyIV).ThesamplesforstudiesV-VIwerefixedinformalinandembeddedinparaffinuntilusedinimmunohistochemistry.

Patients with aseptically loose total hip replacements

IT samples surrounding aseptically loosened femoral stem or acetabular components or their PMMA mantle (studies I-VI) and pseudocapsule (studies I-IV) tissue samples were collected in revision THR operations. All patients had hip pain and radiolucent line between prosthesis components or PMMA mantle and bone. There was no clinical or laboratory evidence of infection as a cause for loosening. Data for the revision THR patients analysed in the studies I-VI is detailed in Table 2.

49

Table 2. Data for the rTHR patients in the studies I-VI. Cases 1-10 were included in the studies I-III. Cases 1, 3, 5, 6 and 8-18 were included in the study IV (cases 11-15 were analysed with RT-PCR and 1, 3, 5, 6, 8-10 and 16-18 with immunostaining). Cases 19-27 were included in the study V. Cases 19, 20 and 22 -29 were included in the study VI.

OA=primary osteoarthritis; CDH=secondary osteoarthritis due to congenital dislocation of the hip; RA=rheumatoid arthritis; Fx=fracture; A=both components; B=only acetabular cup; C=only prosthesis stem; CoCrMo=cobalt-chromium-molybdenum alloy; TiAlV=titanium-aluminum-vanadium alloy; NR=not registered

case diagnosis age sex time torevision

revisedsite

type of prosthesis alloy Fixation

1 OA 79 M 8 B Mathys cup/

Bimetric stemCoCrMo Cementless

2 OA 72 M 4 A Biomet Head-Neck

TiAlV Cementless

3 OA 65 F 9 B Lord CoCrMo Cementless

4 OA 74 F 1 C Wagner CoCrMo Cementless

5 OA 70 M 7 C Müller CoCrMo Cemented

6 OA 74 F 1 C Wagner CoCrMo Cementless

7 OA 66 M 6 B Biomet TiAlV Cementles

8 OA 63 M 3 C Lubinus CoCrMo Cemented

9 RA 78 F 20 A Brunswik CoCrMo Cemented

10 CDH 37 F 4 C Biomet CDH TiAlV Cementless

11 OA 74 F 4 A Brunswik CoCrMo Cemented

12 OA 78 M 9 A Lubinus TiAlV Cemented

13 OA/RA 54 F 10 A Brunswik CoCrMo Cemented

14 OA 73 F 4 A Biomet TiAlV Cementles

15 OA 79 F 14 C McKee-Arden CoCrMo Cemented

16 OA 71 M 4 A Biomet Head-Neck

TiAlV Cementless

17 OA 76 F 23 A McKee-Farrar CoCrMo Cemented

18 OA 80 M 4 A Charnley CoCrMo Cemented

19 OA 79 F 13 C NR NR Cemented

20 Fx 75 M 6 A NR NR Hybrid

21 OA 88 M 16 C NR NR Cementless

22 OA 68 F 16 A NR NR Cemented

23 OA 78 M 6 A NR NR Cemented

24 OA 78 F 5 B NR NR Cemented

25 OA 75 F 7 C NR NR Cementless

26 OA 73 F 4 C NR NR Cementless

27 RA 68 F 17 B NR NR Cemented

28 RA 64 M 12 C NR NR Cementless

29 OA 65 M 16 C NR NR Cemented

50

Control sample patients

Variable control connective tissue samples were used in studies I-VI. In studies I, II andIVnon-inflamedsynovialmembranebiopsiesweretakenfromareaswithoutmacroscopicinflammationduringkneetraumaarthroscopyoperations.InstudiesI and II, samples from ten patients were used. Seven patients were women and three were men and their mean age was 41 years (range 13-70 years). In the study IV, samples from ten patients were used for immunostaining. Six patients were women and four men and their mean age was 31 years (15-57 years). For RT-PCR,samplesfromfivedifferentcontrolpatientswereused.Twowerewomenandthree were men. Their mean age was 52 years (14-70 years). In study III, control connective tissue samples were obtained from 9 patients who had mandibular or maxillaryfracturesfixedwithplates.Thesamplesweretakenfromaroundtheplateduring removal of the plates after fracture healing. Five patients were women and four were men. Their mean age was 38 years (16-63 years). In studies V and VI, control synovial membrane samples were collected from 11 OA patients undergoing primary THR. Eight patients were women and three men and their mean age was 73 years (58-82 years).

Extending the study over multiple years beyond the end of 1990's enabled our research group to use also more sophisticated research methods in addition to immunohistochemistry. For example methods like reverse transcription polymerase chain reaction (RT-PCR) and quantitative real-time reverse transcription polymerase chain reaction (qRT-PCR) were not available to our research group in the beginning of this study.

immunohistochemistry and immunoFluorescence staining

Monoclonal antibodies (MAb) recognizing epitopes of intracellular molecules or cell-surface molecules were used to characterize cells using immunohistochemical staining. Such markers enable the recognition of not only cell type but also the functional activity, e.g. production of activation associated cytokines. The MAbs are categorized according to the epitopes that they recognize according to CD numbers. CD refers to ‘cluster of differentiation’. For example, CD68 is an 110kDa transmembrane glycoprotein present in circulating monocytes and tissue macrophages, which is also present in multinucleated giant cells (MNGCs). Different immunohistochemical staining methods were used in studies I-VI. Primary antibodies used in studies I-VI are listed in Table 3.

51

Table 3. Primary antibodies of the studies I-VI.

antibody source manufacturer

anti-human bFGF, affinity purified IgG rabbit Beckton Dickinson Labware

anti-human TGF-β1, IgG rabbit British Bio-Technology

anti-human TGF-β2, IgG rabbit British Bio-Technology

anti-human PDGF A-chain, IgG rabbit Genzyme

anti-human PDGF B-chain, IgG rabbit Genzyme

anti-human CD68, IgG1 mouse DAKO

anti-human IGF-I, affinity purified IgG mouse Cymbus Bioscience

anti-human IGF-II, affinity purified IgG mouse Cymbus Bioscience

anti-human fibroblast, IgG1 mouse DAKO

anti-human BMP-2, affinity purified IgG goat Santa Cruz

anti-human BMP-4, affinity purified IgG goat Santa Cruz

anti-human BMP-6, affinity purified IgG goat Santa Cruz

anti-human BMP-7, affinity purified IgG goat Santa Cruz

anti-human VEGF, IgG rabbit RDI Research Diagnostics

anti-human VEGF-receptor1, IgG rabbit RDI Research Diagnostics

anti-human VEGF-receptor2, IgG rabbit RDI Research Diagnostics

Avidin-biotin-peroxidase complex (ABC) staining (I-V; for TGF-β1 in study II)

FrozensamplesofrTHRandcontroltissuewerecutto6μmthicksections,mountedongelatinecoatedslidesandfixedin10%bufferedformaldehyde(II)oracetone(I,III-V)for5minat4ºC(I-IV).InstudyV,5μmthickparaffintissuesectionsweredeparaffinizedinxylene,rehydratedinagradedethanolseriesanddistilledwater followed by treatment with 0.4% pepsin in phosphate buffered saline (PBS) in 0.01 N HCl for 30 min at 37º C for antigen retrieval and 3 x 5 min washes in tap water, in 0.01% Triton X-100 surfactant (Sigma-Aldrich Co., St. Louis, USA) in PBS at 22º C and in Tris-HCl buffered saline (TBS). In studies I-IV the antigen retrieval step was not necessary. Endogenous peroxidase activity was blocked with 0.3% hydrogen peroxide (H2O2) in absolute methanol for 30 min at 22º C.

The sections were then incubated in 1) relevant normal serum for 30 min at 22º C; 2) primary antibodies overnight at 4º C; 3) relevant biotinylated IgG for 30 min at 22º C; 4) avidin-biotin-peroxidase complex (ABC) for 30 min at 22º C. TBS or PBS (in study IV) or Triton X-100 in PBS (in study V) was used as diluents. Finally, the sites of peroxidase binding were revealed with a combination of 0.006% H2O2 and 3,3'-diaminobenzidine (DAB). Between two steps the sections were washed 3 x 5 min in TBS or PBS or Triton X-100 in PBS. All sections containing consecutive

52

sections were processed further with and without counterstaining with hematoxylin before dehydration in a graded ethanol series, clearing in xylene and mounting in Diatex (Sigma-Aldrich). Omission of the primary antibody and use of normal serum or isotype matched IgG at the same concentration as and instead of the primary specificantibodiesservedasnegativestainingcontrols(I-V).

ABC - alkaline phosphatase-anti-alkaline phosphatase (APAAP) double staining (III and IV)

The ABC staining was performed as described above (except that in study III, the primary antibody incubation was performed for only 45 min at 22ºC instead of overnight). After treatment of the sections with the H2O2 substrate and DAB chromogen, they were rinsed with tap water and incubated with 1) monoclonal mouseanti-humanfibroblastormacrophage(CD68)antibodiesfor30minat22ºC;2) rabbit anti-mouse IgG for 30 min at 22 ºC; 3) alkaline phosphatase-anti-alkaline phosphatase complex (APAAP) for 30 min at 22º C; 4) bromchloroindolyl phosphate (BCIP) and nitro blue tetrazolium (NBT) for 30-45 min at 22º C in darkness. The bluish colour signal from BCIP/NBT of the APAAP staining clearly differed from the brown colour produced by H2O2/DAB in the ABC staining. Sections were dehydrated in a rising ethanol series, cleared in xylene and mounted in Diatex. Between incubation steps, sections were washed 3 times with TBS.

Peroxidase-antiperoxidase (PAP) staining (for TGF-β2 in study II)

Signal-noiseratiowasbetterforTGF-β2whenPAPstainingwasusedinsteadofABC staining. Incubation times and temperatures were as for ABC staining for TGF-β1,butnormalswineinsteadofgoatserumwasusedtodiminishnon-specificbinding.Rabbitanti-humanTGF-β2IgGwasusedasprimaryantibodiesfollowedby incubations with swine anti-rabbit IgG antibodies and rabbit anti-horseradish peroxidase-horseradish peroxidase (PAP) complexes. Peroxidase binding was revealed with 0.006% H2O2 and DAB as in ABC staining and the rest of the protocol followed ABC staining protocol.

PAP-APAAP (for TGF-β2 in study II)

The PAP staining was performed as described above. After treatment of the sections with H2O2 and DAB, they were rinsed with tap water and incubated with 1) normal

53

rabbit serum for 60 min at 22º C; 2) mouse anti-human macrophage (CD68) IgG for 60 min at 22º C; 3) APAAP solution for 60 min at 22º C. Between incubation steps, sections were washed 3 times with TBS. The rest of the protocol was similar with ABC-APAAP staining.

Streptavidin immunoperoxidase staining (VI)

5μmthickparaffintissuesectionsweredeparaffinizedinxyleneandrehydratedin a graded ethanol series and distilled water and washed 3 x 5 min in Triton X-100 in PBS at 22º C. Then they were treated with 0.4% pepsin in PBS in 0.01 N HCl for 30 min at 37º C for antigen retrieval, followed by 3 x 5 min washes in tap water, in 0.01% Triton X-100 in PBS at 22º C and in TBS. Then the sections were installed in the TechMate Horizon immunostainer (Dako) and stained automatically at 22º C using the following streptavidin immunoperoxidase protocol: 1) polyclonal rabbit anti-human IgG against VEGF; VEGFR1 and VEGFR2 for 90 min at 37º C, diluted in ChemMate antibody diluent (Dako); 2) biotinylated secondary antibody for 25 min; 3) block peroxidase for 25 min; 4) peroxidase-conjugated streptavidine 3 x 3 min; 5) horseradish peroxidase substrate buffer; and 6) substrate working solution containing DAB tetrahydrochloride (ChemMate detection kit) for 5 min. Between steps the sections were washed with ChemMate washing buffers 3 x 5 min. The sections were removed from the machine and one of two sections on each microscope glass slide was counterstained with hematoxylin before mounting. As negativestainingcontrol,theprimaryspecificrabbitIgGantibodieswerereplacedwith non-immune rabbit IgG.

Immunofluorescence staining of cultured cells (V)

Cellswerefixedin4%paraformaldehydefor20minat22ºC,washedwithTriton-Xin PBS for 2 x 10 min, followed by incubations in 1) 5% normal donkey serum in PBScontaining0.1%bovineserumalbumin,for60min;2)affinity-purifiedprimarypolyclonal goat anti-human BMP-2, BMP-4, BMP-6, or BMP-7 IgG antibodies (Santa Cruz Biotechnology) for 60 min; and 3) Alexa Fluor 568-labeled donkey anti-goat antibodies (Molecular Probes, Eugene, OR) for 60 min. Before mounting, the nuclei were stained with 4,6-diamidino-2-phenylindole (DAPI; Vector Laboratories) for 5 min. As negative staining control, non-immune goat IgG was used at the same concentration as but instead of the primary antibodies. The cells were observed underanimmunofluorescencemicroscopeusingappropriatefiltersandcoupledto a digital camera.

54

cell cultures and stimulations

Mesenchymal stem cell culture (V)

Human passage-4 Poietics MSCs (Lot number 6F4392; Lonza, Walkersville Inc., Walkersville, MD) were cultured in Poietics mesenchymal stem cell growth medium according to the instructions of the provider (Lonza). The cells were seeded onto 6-well plates containing cover slips, at 3,500 cells per cm2. For osteogenic differentiation, cells were grown for 14 days in low-glucose DMEM containing 0.1 μMdexamethasone,50μMascorbicacid-2-phosphate,10mMβ-glycerophosphate(Sigma-Aldrich), and 10% foetal calf serum. Differentiation of the MSCs to OBs was checked using AP staining, osteocalcin staining, and staining of the bone mineral as described in detail elsewhere (Myllymaa et al., 2010).

Fibroblast culture and stimulations (VI)

Tissue samples were cut into 3–5 mm3 pieces and explanted onto 25 cm2 tissue cultureflasks(NunclonDelta) inRPMI-1640medium(BioWhittaker®,Lonza)supplementedwith10%FCS(BioWhittaker®,Lonza)and10xantibiotics(1,000IU/mlpenicillinGand1mg/mlstreptomycinsulphate)andkeptinahumidified5%CO2 in air. Next day the medium was changed and the concentration of antibiotics was decreased to 1x. The media was changed twice a week and when 60% of the dish area was covered with a monolayer of cells, the tissue pieces were removed andthecultureswereallowedtogrowtoconfluence.Thecell linesusedintheexperimentshadbeenpassagedfourtofivetimesbydetachingthefibroblastswith0.5% trypsin in Hank’s balanced salt solution (HBSS) without Ca2+ and Mg2+. The cultures were free of mycoplasma contamination as screened with Hoechst 33258 fluorochrome.Inhypoxiastudiescellsfromthreedifferentfibroblastcelllineswereculturedonsix-wellplatestoconfluenceandthenkeptfor8hineitherregular(95% air, 5% CO2) or (N2 controlled) hypoxic (1% O2, 5% CO2) conditions. For cytokinestimulations,cellsfromthreedifferentfibroblastcelllineswereseededin12-wellplates,growntoconfluenceandstimulatedinrespectivemediawithoutorwithinterferon-γ(IFNγ,10ng/ml),interleukin-1β(IL-1β,10ng/ml),interleukin-4(IL-4,10ng/ml), tumournecrosis factor-α(TNF-α,10ng/ml),or interleukin-6(IL-6, 10 ng/ml) for 24 h. All stimulations were conducted in duplicate samples. The cytokines were purchased from R&D Systems.

55

reVerse transcription polymerase chain reaction, rt-pcr (iV)

IGF-IandIGF-IImRNAwas isolated from6μmthickcryostat sectionsusingoligo(dT)25 covalently attached to superparamagnetic polystyrene microbeads via 5' linker group (Dynabeads mRNA DIRECT kit; Life Technologies, Gaithersburg, MD). Cryostat sections from both sides of the sample sections were stained with hematoxylin for histologic examination. Recombinant Thermus thermophilus polymerase (rTth) reverse transcription (Perkin Elmer, Waltham, MA) reaction wasperformedusing5.0UofrTthDNApolymeraseinatotalvolumeof20μl(90mMKCl,10mMTris-HCl,pH8.3,1mMMnCl2,200μMofdATP,dCTP,dGTPanddTTP) on a thermal cycler (Pharmacia, Gene ATAQ Controller, Sollentuna, Sweden). The reaction was run for 1 min at 22º C, 1 min at 37º C, 5 min at 55º C and 10 min at70ºCwithmixingat1-minintervalsafterwhichtheDynabeadassociatedfirststrand was collected using a magnet. The primer design was based on published cDNA sequences and the primer sequences are provided in the original publication V. The second-strand synthesis was performed in PCR buffer containing 5' primer (0.17μMfinal)and100μMofdATP,dCTP,dGTPanddTTPinatotalvolumeof30μl.0.5UofthethermostableDNApolymerase(FinnzymesOy,Espoo,Finland)was added, samples were denatured at 95º C for 2 min and annealed at 55º C / 58º C (IGF-I / IGF-II) for 1 min followed by 5 min extension at 72º C. After melting the strands for 2 min at 95º C, the beads were collected with a magnet and the second-strand cDNA supernatant was transferred to a new PCR tube, where the 3' primer (0.17μMfinal)waspipetted.45cyclesof1minat95ºC,1minat55ºC/58ºC(IGF-I/IGF-II)and1minat72ºCwereperformed.AmplifiedDNAwasrunona1%modifiedagarosegel(FMCBioproducts)forsizeverification.MessengerRNAextractionwascontrolledusingprimersforβ-actin.AnegativeRT-PCRmethodcontrol was run using controls without templates.

QuantitatiVe real-time reVerse transcription polymerase chain reaction, Qrt-pcr (Vi)

TheexpressionlevelsofVEGFanditsreceptors1and2infibroblastculturesinnormoxic and hypoxic conditions and after stimulation by various cytokines was measured using qRT-PCR. Total RNA from the cultured cells was isolated using RNeasy Mini Kit (QIAGEN) according to the manufacturer’s instructions. 500 ng of total RNA was reverse transcribed using iScript cDNA Synthesis Kit (BioRad Laboratories).QuantitativeRT-PCRwasrunusing10ngfirst-strandcDNAand0.5 mM primers in iQ SYBR Green Supermix (Bio-Rad) in iCycler iQ5 Multicolor

56

Real-Time PCR Detection System (Bio-Rad). Primers were designed with Primer3 (SourceForge: http://primer3.wi.mit.edu), the sequences were searched using the NCBI Entrez search system and sequence similarity search was done using the NCBI Blastn program. The primer sequences are provided in the original publication VI. The copy numbers of mRNA were determined in duplicates and normalized against β-actinandporphobilinogendeaminasegenes.

QuantiFication oF immunohistochemical staining

Semi-quantitative image analysis (I-IV)

The average number of positive cells from immunohistochemical staining was calculatedinfivedifferenthigh-powerfields(x400)usingalowlightcharge-screencoupled CCTV camera (Panasonic WV-CD 130L) mounted on an Olympus BH-2 light microscope linked to a semiautomatic Kontron image analysis and processing system (Kontron Bildanalyse) equipped with the VIDAS 2.1 program (Kontron Elektronik). The number of positive cells was calculated in three representative samples in all study groups to get an estimate for a population mean. Using the formulat=(x-µ)/SEMitwascalculatedthatthemeanoffivehigh-powerfieldsisnotsignificantly(p>0.05)differentfromthepopulationmean(treferstot-statistics,x=the sample mean, µ=the population mean, SEM=standard error of the population mean). Results were expressed as the number of positive cells per 1 mm2 of tissue.

Crude visual analysis (V-VI)

Stained tissue sectionswere analysedunder400×magnification.Microscopicfindingsweregradedasfollows:0=noimmunoreactivecells;1=onlyoccasionalimmunoreactive cells (< 10%); 2 = moderate numbers of immunoreactive cells (10–50%); 3=many immunoreactive cells (> 50%). Based on their location,arrangement,andmorphology, the immunoreactivecellswereclassifiedinto1)synovialliningcells,2)spindle-shaped,stromalfibroblast-likecells,3)interstitialmonocyte/macrophage-like cells, and 4) vascular endothelial cells. Intensity of the staining was evaluated separately. The evaluation was also based in part on our experienceintheidentificationofsuchcellsthroughourextensivemarkerstudiesof interface tissues (Santavirta et al., 1990a; 1990b; 1998).

57

statistical analysis

In studies I-IV for semi-quantitative image analysis results, statistical BMDP-PC 7.01 software was used to calculate the mean and standard error of the population meantodescribethedispersionofthedata.Thesignificanceofdifferencesbetweenmeans was analysed using Mann-Whitney U test for skewed or Student’s t-test for normally distributed variables. A p value of 0.05 or less was considered statistically significant.

In studies V-VI for crude visual analysis results, statistical SPSS software for Windows version 11.5 was used to calculate the medians and 95% CIs for medians of cell density scores. Mann-Whitney U test was used for comparison of paired variables between revision THR and OA groups, because the data had an ordinal level of measurement and was therefore non-parametric. Wilcoxon signed rank test was used for comparison between variables within the rTHR and control groups using 2-tailed tests.

58

5 RESULTS AND DISCUSSION

General overview of the results in the studies I-VI is presented in Table 4.table 4. General overview of the results.* trend with no statistical significance** statistically significant difference when compared to control synovial tissue

interfacevs. control

pseudocapsulevs. control macrophages fibroblast-

like cellsendothelialcells

bFgF +++ ++ +

tgF-β1 +++ ++ +

tgF-β2 +++ ++ +

pdgF-a +++ ++ ++

pdgF-B +++ ++ ++

igF-i + + +

igF-ii similar +++ +++ ++

Bmp-2 similar not studied ++ + ++

Bmp-4 * not studied +++** ++ ++

Bmp-6 similar not studied +++ +++ ++

Bmp-7 similar not studied ++ + +

VegF similar not studied + + +

VegFr1 similar not studied ++** ++** ++

VegFr2 * not studied + +** +++

histopathological eValuation oF interFace tissue and pseudocapsular tissue

Sample tissue obtained at the time of the revision of THR was synovial-like membrane. It was covered with a synovial lining-like layer, usually one to ten cells in thickness. Subsynovial vascularised stroma contained mainly macrophages/foreign bodygiantcells,oftenorganisedtogranulomas,andfibroblasts.Highlycellularareascontainedmostlymacrophagesandfibroustissueconsistedmainlyofrelativelydensecollagenousfibresandelongatedfibroblasts.Ingeneral,thefibroustissueoccupied mainly the deeper bone side of the membrane, whereas the macrophage-rich tissue was located closer to the synovial-like lining layer. Morphologically IT and pseudocapsular tissue were very similar. There were no apparent structural differencesdependingon the typeoffixation(cementedversuscementless)or

59

prosthetic materials (CoCrMo versus TiAlV). The synovial-like membrane of IT and pseudocapsule was very similar to synovial membrane from control hip or knee synovial membrane. The main difference was the lack of debris and foreign body reaction in the control synovial tissue.

wear particulate deBris

Particulate debris in IT and pseudocapsular tissue included UHMWPE, PMMA, and metalweardebris.LargeUHMWPEparticleswereidentifiedashighlybirefringentfinefragmentswhenviewedunderpolarized light.PMMAparticlesareusuallydissolved by solvents (xylene) used for sample processing. PMMA particles were usually “seen” as large and round empty spaces surrounded by connective tissue and macrophages / foreign body giant cells. Occasional undissolved PMMA-like particleswerealsoseen.Metalparticleswereidentifiedasirregularlyshaped,fineblack particles. Varying numbers of small metal particles had been phagocytosed by macrophages / foreign body giant cells, but particulate debris was also frequently seen in the extracellular tissue. However, the chemical nature of these particles could not be ascertained. Perhaps some of these particles represent radiocontrast agents like BaSO4, ZrO3orbone.Identificationandsystematicclassificationofindividualparticles as UHMWPE, PMMA, and metal debris was not done in this study.

Basic FgF in aseptic loosening

InstudyI,bFGFwasfirsttimereportedtobepresentintheITandpseudocapsulartissue samples which were obtained at revisions performed for aseptic loosening ofTHR.Itwasfoundparticularly in interstitialfibroblastsandinmacrophages,which were the predominant cells in these patient samples. Weaker immunostaining was found in less numerous vascular endothelial cells of capillary blood vessels and post-capillary venules. Basic FGF positive macrophages seemed to be located predominantly in cell-rich areas, which coincided with implant derived debris, either UHMWPE,PMMAormetalparticles.BasicFGFpositivefibroblast-likecellsweremore evenly scattered. The number of bFGF positive cells was relatively similar in IT and pseudocapsular tissues, cemented and cementless hips and CoCrMo and TiVlA alloy implants. However, patient numbers were too low for proper statistical comparison.

Control samples of synovial membrane obtained at the time of the knee arthroscopies were structurally very similar with a layer of synovial lining cells and vascularised sublining connective tissue. The main difference in the control tissue

60

samples(I)wasthelackofdebrisandinfiltratingmonocyte/macrophage-likecellsgranulomas. Image analysis showed that the numbers of bFGF positive cells were higher both in the IT (1693 ± 291 cells/mm2 tissue) and pseudocapsular tissues (1954 ± 256 cells/mm2 tissue) than in the control synovial membranes (1009 ± 133 cells/mm2 tissue, p<0.05 and p<0.01, respectively).

Concurrently with study I, an in vitro cell culture study showed that a low concentrationofTiAlVparticlesstimulatedhumanforeskinfibroblaststoproducemorebFGFthanhigherconcentrations(Manlapazetal.,1996).Later,ourfindingwasconfirmedandbFGFexpressionwasfoundtobeelevatedinPTsamplescomparedwith normal synovial tissue samples (Koreny et al., 2006). The same group described also that similar amounts of bFGF were produced by IT and normal synovial tissue explants samples in vitro as analysed by RNase protection assay (Tunyogi-Csapo etal.,2007).Thisseemstobesomewhatinconflictwiththeirearlier(Korenyetal.,2006)andourfindings(I).Maybethisinpartrelatestothetraumatictissueexplants sample collection, because the amount of bFGF in IT samples increased significantlyaftertissuecultureforoneweek.Furthermore,ITfibroblastswerehighly induced to produce bFGF, when they were treated with culture media from IT,withorwithouttitaniumparticlesortreatedwithaninflammatorycytokinecocktail.Thisisinconcordancewithourfindings(I).

To sum up, bFGF has been studied only in a few studies in aseptic loosening sofar.WearparticlesseemtoinducebFGFproductionespeciallybyfibroblastsand macrophages. The expression of bFGF in PTs is at least not diminished but mostprobablyincreasedinasepticloosening.Thefindings,togetherwithcurrentknowledge of the many roles of bFGF, suggest that it helps in the growth and maintenance of periprosthetic connective tissue and its vasculature as a response to wear debris. Increased amount of bFGF could have positive effects on periprosthetic bone.Thismayhoweverbeinsufficienttocompensatefortheevenmoreincreasedosteolysis. Furthermore, the effects of bFGF are pleiotropic and it has also been shown to inhibit OB differentiation and function and to be able to stimulate OC recruitment and activation via RANKL, in association with physiologic and pathologic bone remodelling (Collin-Osdoby et al., 2002; Shimoaka et al., 2002). Finally, the increased number of bFGF+ cells per tissue area might also be explained bythehighercellnumbersintheinflamedPTscomparedtocontroltissues.Thesethree explanations can also be applied to some of the other cytokines described below.

61

tgF-β1 and tgF-β2 in aseptic loosening

TGF-βmRNAwasfirstdetectedinPTsbyinsituhybridizationwhilestudyIIwasunderway,butthecellularsourcesofTGF-βwerenotdetermined(Goodmanetal.,1996).Ataboutthesametime,TGF-β1wasfoundinconditionedmediafrom24hcultured tissue explants samples of periprosthetic pseudosynovial membranes, but theamountofTGF-β1wasnotrelatedtoradiologicalfindingsoffailedimplants.Furthermore,nosignificantcorrelationswerefoundbetweenlevelsofTGF-β1andIL-1β,IL-6,TNF-αorPGE2(Perryetal.,1997).StudyIIconfirmedthesefindingsandextendedtheresultstoTGF-β1andTGF-β2isoformsbyimmunohistochemicalstainingwithPAP-APAAPdoublestainingandverifiedmacrophagesasthemaincellularsourceofTGF-βsinPTs.

InstudyII,TGF-β1andTGF-β2stainingpatternwasstrikinglysimilartothatofthebFGF(I).PositivelystainedmacrophagesandfibroblastsweremostnumerousoftheTGF-β+cellsandalsodisplayedthehigheststainingintensity,whilethenumberandstainingintensityofECswasclearlylower.AfewTGF-β+lymphocyteswerealsonoticed.PAP-APAAPdoublestainingconfirmedthatmacrophageswerethemainsourceofTGF-β1andTGF-β2positivecells(II).AswithbFGF(I),thestainingresultsweresimilarregardlessofimplantalloy,implantfixationtypeandsamplesite.EvenstainingfindingsofcontrolsynovialmembraneresembledthoseofbFGF(I).ImageanalysisshowedthatthenumberofTGF-β1and–β2positivecellswerehigher both in ITs (2327 ± 212 and 2292 ± 594 cells/mm2 tissue; respectively) and pseudocapsular tissues (1796 ± 214 and 3210 ± 585 cells/mm2 tissue) than in the control synovial membranes (946 ± 136 cells/mm2 and 311 ± 113 cells/mm2 tissue, respectively; all p values <0.01) (II). As with bFGF, the cellular origin was similar in all tissues (I,II).

In 1996, slightly before study II was published, it was reported that human foreskinfibroblastsexposedtoTiAlVparticlesatdifferentconcentrationsdidnotproduceTGF-βatall(Manlapazetal.,1996).Inanotherstudy,TGF-β1expressionof human peripheral monocytes was found to be inhibited by titanium, chromium and cobalt ions (Wang JY et al., 1996a). Furthermore, PMMA or CoCr particles did notaffectTGF-β1mRNAexpressionofmurinemacrophages(Prabhuetal.,1998).AlltheseinvitrofindingsseemclearlytobeinconflictwithourfindingsinstudyII.WefoundincreasedTGF-β1andTGF-β2positivecellcountinPTsofasepticallyloosened hip implants when compared to control knee synovial membrane samples andtheTGF-β+cellsweremainlymacrophagesandfibroblasts(II).Thismayinpartrelate to the short term stimulation in in vitro studies compared to the often 10-15 year long process in PT samples before rTHR becomes necessary. Second, these types of in vitro studies are biased due to the lack of other cells, tissue architecture and lack of multiple cellular interactions, which can occur in vivo, such as continuous

62

recruitmentofcirculatingcellsandinteractionsbetweenmacrophages,fibroblasts,vascular ECs, mast cells and many others. For both of these reasons, processes like developmentoffibroticimplantcapsuleorsynovial-likeliningcannotbeseeninvitro.

Later,similarlevelofTGF-β1mRNAwasfoundusingdensitometricallyanalysedRT-PCR in ITs retrieved at hip revision arthroplasty as in control samples of OA and rheumatoid arthritis patients (Nabae et al., 1999). This may in part relate to thenormalisationofTGF-β1mRNAagainstthehousekeepinggene:TGF-βmaybeincreasedinPTsduetoeitherincreasedproductionofTGF-βpercellorincreasednumbersofTGF-β+cellspertissuearea(orvolume).DuetothehighsensitivityofRT-PCR, relatively small sample volumes are used, which may lead to a bias due to the tissue sampling area (i.e. where exactly the sample was taken). Finally, it is nowadayswellrecognizedthatregulationoftheTGF-βproteinproductionmayalso occur at the post-transcriptional level (translation, mRNA stability etc.) so that the mRNA and protein level studies do not always agree with each other. At thesametime,inlinewithourwork,TGF-βimmunoreactivitywaslocalisedinPTtofibroblastsandmacrophagesaswellasinperiprostheticbonetoOBsandbonematrix(Al-Saffaretal.,1999).TheauthorsdidnotreportTGF-βimmunoreactivityinendothelial cellsor in inflammatory cell infiltrateswhereaswe founda fewendothelialcellsandoccasionallymphocytestobeTGF-βpositive(II).However,they foundthatTGF-βwashighlyexpressedat thesitesofosteogenesiswhichsuggestsanactiveroleforitinboneregenerationandtransitionfromfibroustissueto bone (Al-Saffar et al., 1999).

InaccordancetostudyII,theamountofTGF-β1protein,measuredwithhigh-throughput protein chips, has been found to be more than threefold higher in homogenized PT samples than in homogenized OA synovial capsular tissue samples (Shanbhagetal.,2007).HigheramountsofTGF-β1hasbeendescribedinfreshIT explants samples than in normal synovial tissue samples by RNase protection assay(Tunyogi-Csapoetal.,2007).TheamountofTGF-β1increasedsignificantlyafter tissuecultureforoneweek.Furthermore, itwasfoundthatITfibroblastswereinducedtoproduceTGF-β1whentheyweretreatedwithconditionedculturemedia from interface membrane (CM-IFM), with or without titanium (Ti) particles ortreatedwithaninflammatorycytokinecocktail.Interestingly,TiparticlesaloneinducedhigherTGF-β1expressionthantreatmentwithCM-IFMwithorwithoutTi particles, which differed dramatically from the additive response of Ti particles and CM-IFM to bFGF and VEGF expression (Tunyogi-Csapo et al., 2007).

THRmaterialsmayaffectTGF-β levels inPTs, since failedmetal-on-metalTHRswerereportedtobeassociatedwithfewerTGF-βpositivecellscomparedwith failed metal-on-PE THRs (Campbell et al., 2002), which might in part be due to the direct toxic effects of metal ions on cells. In vitro, 0.5 – 3.0 µm HA particles inhibitedTGF-β1productioninOB/OCco-cultureswhereas largerparticlesdid

63

nothaveanymarkedeffect(Sunetal.,1999).Likewise,≤1µmsizedUHMWPE(Dean et al., 1999) and 1 µm sized Al2O3 (Lohmann et al., 2002) particles decreased TGF-βlevelsofOBcultureconditionedmedia.Furthermore,1–10µmPMMAparticles inhibitedosteoblasticexpressionofTGF-β1byday4butnotbyday8inacellculturestudy(Maetal.,2010).Duetomatrixbinding,TGF-βmightplaytissueandcellspecificrolesindifferentlocations.LowerconcentrationsofTGF-β1havebeenfoundinthejointfluidsfrompatientswithloosekneeorhipimplantscomparedtopatientsundergoingfirstimplantation,butserumTGF-βlevelsdonotseemtochangesignificantlyinasepticloosening(WangJYetal.,1996b;Fioritoetal.,2003).Inonestudyhowever,plasmalevelof latentTGF-β2waslowerinpatients with implant loosening than in patients with stable hip implants. In the samestudy,thelevelsofHCl-activatedTGF-β2andlatent/activatedTGF-β1weresimilar in both patient groups (Cenni et al., 2005). It is not known from what tissuecompartmentthejointfluidandserumTGF-βisderivedoriftheassaysusedonlymeasuredbiologicallyactiveTGF-βoralsolatentcomplexesand/orinactivedegradationproductsofTGF-β.Finally,itmaybethatlocalactivationofTGF-βleadstohighaffinitybindingtospecificreceptors,preventingdiffusionawayfromthe site of activation.

Inconclusion,TGF-β-relatedfindings inaseptic looseningseemtoat leastpartially contradictory. However, the majority of the studies since study II show theexpressionofTGF-βtobeincreasedinvitroinfibroblastsandmacrophagestreatedwithwearparticles.Correspondingly,theexpressionofTGF-βisincreasedinfibroblastsandmacrophagesininflamedtissueareasofthePTsinasepticloosening.

TGF-β ismost likelynecessary in inflammatoryreactions,angiogenesisandconnective tissue growth and remodelling of PT surrounding loosening implants. Eventhoughithasosteolyticproperties,TGF-βprobablyhasmainlypositiveeffectson periprosthetic bone in aseptic loosening since it is highly expressed at the sites of bone repair/formation (Al-Saffar et al., 1999).

pdgF in aseptic loosening

PDGF-B(alsoknownasPDGF-2)mRNAwasfoundbyinsituhybridizationfirsttimeinPTmacrophagesandfibroblastsbyJiraneketal.(1993).Later,PDGF-A(alsoknownasPDGF-α)mRNAwasdetectedbyinsituhybridizationinPTs,butthecellularsourceswerenotdetermined(Goodmanetal.,1996).StudyIIIconfirmedthesefindingsbyimmunohistochemicalstainingandprovidedmoreinformationoncellularsourcesofPDGFswithquantificationofpositivelystainedcells.

PDGF-A and PDGF-B chain containing cells were found in all interface and pseudocapsular tissues from cases of aseptic loosening of THR. Staining pattern

64

resembledthatofbFGFandTGF-βssincePDGF-AandPDGF-Bchainswerefoundinmacrophages,fibroblastsandvascularendothelialcells.AswithbFGFandTGF-βs,macrophagestainingwasstrongandmacrophagewasclearlythemostdominantPDGF+celltypeinallsamples.ThiswasalsoconfirmedinanABC-APAAPdoublestaining using PDGF and CD68 antibodies. About half of CD68 positive cells were PDGF-A and PDGF-B positive, often containing implant debris. The PDGF-A and PDGF-B positive cells in similar cell types were also seen in control tissues. Perhaps due to small patient numbers, differences of PDGF immunostaining between differentimplantalloysorfixationtypesdidnotreachstatisticalsignificance(III).Image analysis showed that the numbers of PDGF A and B chain positive cells around loose THR in IT (1881 ± 486 and 1877 ± 214 cells/mm2 tissue) or pseudocapsules (1786 ± 236 and 1676 ± 152 cells/mm2 tissue) were higher (p<0.01) than in control tissues (821 ± 112 and 467 ± 150 cells/mm2 tissue), respectively (III).

Regardingtheeffectofwearparticlesonfibroblast/macrophagePDGFexpression,ithasbeenreportedthathumanforeskinfibroblastsexposedtosubmicronsizedTiAlV particles at different concentrations, in an enzyme-linked immunosorbent assaystudy(ELISA),didnotproducePDGF-AB(orTGF-β),butproducedbFGF(Manlapaz et al., 1996). In another ELISA study, submicron sized TiAlV particles did not increase the PDGF production by cultured human macrophage-like cells (Rolfetal.,2005).Thesefindingsseemtobeinconflictwithourfindings(I),butcells in periprosthetic tissue may behave differently to cultured cells and the nature of cytokine response in aseptic loosening is known to depend on the composition, size, shape, volume, and surface area of the particulate debris (Purdue et al., 2007). Biomechanicalloading,fluidshearstress,hypoxia,pHandcytokinenetworkeffectsmay play a role in vivo. Indeed, it is currently believed that particles alone cannot explain the so called particle disease, but requires co-stimulation, e.g. particle opsonisation with danger associated molecular signals (Gallo et al., 2012). No differences have been detected in PDGF-AB or PDGF-BB plasma levels between patients with loosened and stable hip implants (Cenni et al., 2003; 2005). It is possible that locally released PDGF binds to cell receptors in the PT or is degraded by activated proteases and thus does not reach (enter) the blood stream. Our group hasproposedthatthereis insufficientdegradationofPDGF-receptorcomplexesin aseptic loosening, which could lead to the accumulation of connective tissue (Niissalo et al., 2002).

As described above, PDGF has not been studied extensively in aseptic loosening. It plays a role in theupregulationof tissuemetabolismboth in inflammationand wound / connective tissue healing. Its presence is presumed to precede or coincidewiththestimulationofothercytokines,suchasTGF-β.Itisalsoknownthat PDGF can stimulate both bone formation and resorption. It is important to recognizethatnotonlymacrophagesbutalsofibroblastsandvascularendothelial

65

cells produce PDGF. Because of the variety of cell types present in PT, it seems likely that multiple potential mechanisms of PDGF action exist in aseptic loosening. It probably assists in periprosthetic connective tissue growth, angiogenesis and remodelling. Furthermore, it may act as a chemokine and mitogen for macrophages, fibroblastsandosteoblasts.Itmayalsoenhanceboneformationthroughregulationof bone remodelling or via direct effects on osteoblasts. Another, completely opposite possibility is that PDGF stimulates bone resorption directly or indirectly via, for example, stimulation of prostaglandin-E2, which is highly expressed in IT.

igF-i and ii in aseptic loosening

StudyIVwasthefirstonetoelucidatetheroleofIGF-Iand-IIinasepticloosening.RT-PCRrevealedthatbothIGFswereproducedinallfivestudiedrTHRandfivecontrol knee synovial samples. IGF-I and -II immunoreactivity was observed in all the studied rTHR and control samples. Interestingly, the cellular sources of the IGF isotypes were somewhat different. Most of the IGF-I containing cells were vascular endothelial cells and monocyte/macrophages. The same cell types werealsoIGF-IIpositive.Inaddition,fibroblast-likecellsshowedstrongIGF-IIstaining.ABC-APAAPdoublestainingwasusedtoconfirmthatbothfibroblastsand macrophages produce both IGF isoforms. Staining pattern was consistent with cytoplasmic rather than membrane-bound localization of IGF immunoreactive material (IV), perhaps suggesting actively ongoing IGF synthesis. The number of IGF-I immunoreactive cells per mm2 was low in the IT samples (295 ± 37) and pseudocapsule (339 ± 72) compared to control synovial tissue (866 ± 123; all p values <0.01). Likewise, the number of IGF-II immunoreactive cells per mm2 was significantlylowerinthepseudocapsule(296±35)comparedtocontrols(602±108; p<0.01). The IGF-II positivity was similar between the controls and the IT (602 ± 108 cells/mm2 vs. 596 ± 97 cells/mm2, NS). Although the numbers were too low for proper statistical comparison, there seemed to be no difference in IGF expressionbetweencasesofcementedorcementlessfixationorimplantsmadeof TiVAl or CoCrMo alloy.

In addition to study IV, only one study has covered IGF in aseptic loosening: lowerlevelsofIGF-I,measuredbyradioimmunoassay,werefoundinsynovialfluidand serum of patients with aseptic loosening compared with OA patients. The levels ofIGF-IIweresimilarinthesynovialfluidsamplesofbothstudygroupsandtheserum levels of IGF-II were not studied or reported (Andersson et al., 2005). These resultsresembleourfindings(IV).AmoretrivialexplanationcouldbethatIGFisincreased in OA, e.g. as a mediator in subchondral bone thickening and formation of osteophytes (Middleton et al., 1995). However, the control samples used in study

66

IV were not PTs from primary THR performed for OA but samples of synovial membrane taken during arthroscopy of knee trauma patients.

To sum up, IGFs have been studied very little in aseptic loosening. IGFs are the only growth factors studied in this thesis work that were clearly down-regulated in PTs when compared to the control samples. In contrast, a plethora of locally up-regulatedpro-inflammatorycytokineshavebeenidentifiedinPTs.Probablysome of these cytokines, local tissue conditions or systemic regulatory mechanisms inhibit local IGF production in PTs surrounding aseptically loosened prostheses. Forexample,mitogenicgrowthfactorslikebFGF,TGF-βandPDGFcandecreaseosteoblastic IGF-I and -II secretion (Canalis et al., 1993; Gabbitas et al., 1994) and are up-regulated in aseptic loosening (I-III).

Down-regulation of local IGF production most probably has negative effects on periprosthetic bone formation and on the growth, angiogenesis and remodelling of periprosthetic soft tissues in aseptic loosening. IGFs have been found to induce and even substitute RANKL and thus to enhance OC differentiation, activation and function (Mochizuki et al., 1992; Hemingway et al., 2011), but this effect may not be important in aseptic loosening because PTs are characterized by a high RANKL/OPG ratio (Mandelin et al., 2003).

Bmps in aseptic loosening

Localization and distribution of BMPs in PTs of aseptically loosened hip implants wasfirstdescribedinstudyV.LikeinstudiesI-IVandVIperiprostheticbonewas not studied. Similarly to growth factors in studies I-IV and VI, BMPs werefoundinsynovialliningorlining-likelayers,fibroblast-likestromalcells,interstitial monocyte/macrophage-like cells, and endothelial cells in all samples. The staining patterns were relatively similar in control OA and rTHR samples, but the density of BMP-4 positive monocyte/macrophage-like cells was higher in rTHR samples than in OA samples (p<0.05). The staining intensity of the BMP-4 positive monocyte/macrophage-like stromal cells was also stronger in rTHR samples than in control samples (V).

In rTHR samples BMP-6 positive cells seemed to be generally more frequent than cells that were positive for other BMPs. BMP-2 and BMP-4 positive cells were usually present at intermediate frequencies and BMP-7 positive cells were the least frequent. BMP-6 positive cells were more frequent than BMP-7 positive cells amongfibroblast-likestromalcells(mean2.6vs.0.9,p<0.05),macrophage-likecells (mean 2.9 vs. 1.6, p<0.05) and endothelial cells (mean 2.2 vs. 1.3, p<0.05). Furthermore, BMP-4 positive cells were more frequent than BMP-7 positive cells amongfibroblast-likestromalcells(mean1.9vs.0.9,p<0.05)andmacrophage-like

67

cells (mean 2.4 vs. 1.6, p<0.05). Any other differences between the staining results werenotstatisticallysignificant.

In control OA samples BMP-6 positive cells were more frequent than cells positive for BMP-2, BMP-4, and BMP-7 among all cell types (p<0.05) but not in synovial lining cells between BMP-6 and BMP-2 expression (p = 0.06). There were nostatisticallysignificantdifferencesbetweenthefrequenciesofBMP-2,BMP-4,andBMP-7positivecells.Inconclusion,theonlystatisticallysignificantfindingbetween rTHR and control samples was that BMP-4 positive macrophage-like cells were more frequent in rTHR than in control samples (V).

Expression of BMP-2, -4, -6, and -7 was detected in human bone marrow-derivedMSCsbyimmunofluorescencestainingasgranular,cytoplasmicstaining.The staining diminished during osteogenesis so that osteoblast-rich cell cultures contained less cytoplasmic BMP-immunoreactive granules than the undifferentiated MSC progenitors. This suggests that in particular OB progenitor cells produce BMPs, probably for storage in the bone matrix (V).

ResultsofthestudyVareinaccordancewithapreviousfindingthatpolyethylene-stimulated human macrophage-like cells produce BMP-4 mRNA, detected with RT-PCR (Rader et al., 2002). Two other particle studies were published at the same timewithstudyV.Inthefirstone,1–10µmPMMAparticlesinhibitedosteogenicdifferentiation of MC3T3-E1 osteoprogenitor cells, but addition of BMP-7 counteracted this effect by stimulating the osteogenic differentiation (Kann et al., 2010). In the second in vitro study of the same research group, 1 – 10 µm PMMA particles inhibited the osteoblastic expression of BMP-2 and -4 (Ma et al., 2010). Causal connections are naturally far more complex in real in vivo aseptic loosening situation with mechanical loading, local hypoxia and a plethora of interacting cytokines, but the results suggest that BMPs may be used to slow or reverse aseptic osteolysis.

In conclusion, increased BMP-4 probably results from particle-induced activation of macrophages. It likely affects positively on bone formation but clearly not enough to maintain implant osseointegration. Like in studies I-IV, other interpretations are also possible: BMPs, including BMP-4 can induce osteoclastogenesis by increasing RANKL production by OBs and promote OC survival and they have been found to stimulate OC activation and function (Mishina et al., 2004; Canalis, 2009), except for BMP-7, which inhibits osteoclastogenesis (Maurer et al., 2012). It could be that the relatively high degree of BMP-4 expression in revision THR interface samples is in part responsible for the enhanced activity of osteoclasts, and in fact stimulates peri-implant osteolysis rather than osseointegration. BMPs have been implicated in many diverse responses and their activity is affected by their concentration, other growth factors/ cytokines and mechanical stimuli (for some examples, see Cunningham, 1992; Zeisberg et al., 2003; Csiszar et al., 2008; Meynard et

68

al.,2009).Therefore,ourfindingsmightalsobeexplainedbyconnectivetissuegrowth and by BMP-driven formation of interface tissues. Since BMP-2 and -7 are the only commercially available growth factors in orthopaedic use at the moment, it is surprising that BMPs have not been studied more in aseptic loosening.

VegF and its receptors 1 and 2 in aseptic loosening

First VEGF-related studies in aseptic loosening focused on angiogenesis, macrophages andfibroblasts.TheydescribedthatVEGF,VEGFR1andVEGFR2areexpressedinPTs (Jell and Al-Saffar, 2001; Miyanishi et al., 2003; Spanogle et al., 2006; Koreny et al., 2006; Tunyogi-Csapo et al., 2007), but the results for expression of VEGFR2 werecontradictory.Furthermore,VEGF/VEGFRpositivecellswerenotquantifiedin the papers. Therefore, we decided to systematically quantify all the main VEGF/VEGFR positive cells in PTs. Macrophages are considered key players in aseptic loosening,butalsoresidentfibroblastscanproduceandreleaseVEGFandmaybeactive in aseptic loosening, perhaps even in actively resorbing periprosthetic bone (Pap et al., 2003; Koreny et al., 2006; Tunyogi-Csapo et al., 2007). That provided inspirationtoclarifytheVEGF/VEGFRsystemanditsregulationinfibroblastsbyhypoxiaandpro-inflammatorycytokines,bothpresentinasepticloosening.

VEGF, VEGFR1 and VEGFR2 in rTHR and in OA and in different cells (VI)

The immunostaining of VEGF+ cells was only weak and similar in revision THR and OA samples. There were slightly more VEGF+ cells in all the main cell types (liningcells,macrophage-likestromalcells,fibroblast-likestromalcells,andECs)in rTHR samples than in OA samples, but the differences were statistically not significant.However,VEGFR1wasup-regulatedinsynoviallining-likelayercells,instromalmacrophage-likecellsandespeciallyinstromalfibroblast-likecellsinrTHRcompared to OA samples (p < 0.01; p < 0.05 and p = 0.001; respectively). VEGFR2 wasslightlyup-regulatedinstromalfibroblast-likecells(p<0.05),butnotinanyother cell subtypes. Both in OA (p<0.01) and in rTHR samples vascular endothelial cellscontainedmorefrequentlyVEGFR2thanVEGFR1.VEGFR1+fibroblast-likecellsweremorenumerousthanVEGFR2+fibroblast-likecellsinrTHR(p<0.05),but not in OA samples (VI).

The results of the study VI strongly point towards up-regulation of VEGF/VEGF-system in aseptic loosening. Noteworthy, OA synovial samples were used as controls and VEGF has been shown to be strongly up-regulated in OA synovium when compared to normal synovium (Giatromanolaki et al., 2003).

69

Effects of hypoxia and pro-inflammatory cytokines on VEGF, VEGFR1 and VEGFR2 (VI)

Synovialfibroblastscultured inhumidified5%CO2inairexpressedVEGFandupon8hourexposureofthefibroblaststohypoxiatheexpressionlevelrouse2.4-fold(p<0.01).Incontrast,hypoxiadidnotsignificantlychangetheVEGFR1orVEGFR2expressionlevels.24hourculturewithIFN-γ,IL-1β,TNF-α,andIL-6didnotaffectVEGF,VEGFR1orVEGFR2expressioninfibroblasts.However,IL-4increased the VEGFR1 expression over 20-fold without affecting VEGF or VEGFR2 expression (VI).

InthefirststudyonVEGFexpressioninrTHR,VEGFandVEGFR2positivityhas been shown in macrophages, synoviocytes, ECs and in multinucleated and spindle-shaped cells (Jell and Al-Saffar, 2001). The authors suggested that wear debrisandforeignbodyinflammationdisturbangiogenesisinperiprosthetictissues,since poorest vascularisation was noticed in areas with high levels of wear debris (JellandAl-Saffar,2001).ThefindingscorrespondwiththeresultsofthestudyVI.

Miyanishi et al. (2003) detected expression of VEGF and VEGFR1, but not VEGRF2 in monocyte/macrophages of rTHR tissue samples. However, no VEGF, VEGFR1 or VEGFR2 was detected in OA synovial tissue samples, which is in clear contradiction with the results of earlier studies (Paavonen K et al., 2002; Giatromanolaki et al., 2003) and with the study VI. Titanium particles induced in vitro VEGF expression in macrophages and induced macrophage chemotaxis, which could be inhibited by VEGF antibody (Miyanishi et al., 2003).

OurfindingsconfirmtheseearlierobservationsonthepresenceofVEGFandVEGFR1 containing macrophage-like cells in the IT around loosened implants, and suggest that in particular VEGFR1+ macrophages play a role here (VI). Earlier studies focused on monocyte/macrophages and did not describe what other cells, e.g.vascularendothelialcellsorfibroblasts,expressedVEGForVEGFR1.StudyVIextendstheseearlierfindings,describesVEGFanditstype1and2receptorsinvarious cell types and locations, and compares their presence in PT and OA synovial membrane samples (VI).

Higher amount of VEGF coding mRA molecules have been described in fresh IT samples than in normal synovial tissue samples by using RNase protection assay (RPA) (Spanogle et al., 2006). VEGF expression was higher in PTs than in OA when measuredbyimmunofluorescence,butper-cellmRNAproductionofVEGFwassimilar when measured by RT-PCR. The reason was higher overall and VEGF+ macrophage cell density (Spanogle et al., 2006).

CulturedPTfibroblaststreatedwithconditionedmediafrominterfacemembranes(CM-IFM) with or without titanium particles, expressed higher amount of VEGF comparedtoVEGFexpressioninuntreatedPTfibroblastcultures(Korenyetal.,2006). In a later study of the same group and with the same RPA method, higher

70

amount of VEGF was noticed in fresh and cultured IT samples than in normal synovial tissue samples. When the tissues were cultured for one week, the amount ofVEGFbecameevenmoreincreasedinITsamples.Furthermore,ITfibroblastsproduced VEGF when treated with CM-IFM with or without titanium (Ti) particles ortreatedwithan inflammatorycytokinecocktail.SimilarlytobFGF,CM-IFMand Ti particles had an additive effect on VEGF expression. (Tunyogi-Csapo et al., 2007).AllthesefindingsareinaccordancewiththestudyVI.

A couple of wear particle studies have been conducted covering VEGF in aseptic loosening. Macrophages treated with either highly cross-linked polyethylene (HXPE) or ”conventional” UHMWPE particles secreted increased amounts of VEGF (Illgen et al., 2008). Similar results were achieved with human OBs as their expression of VEGF was also induced by PMMA, ceramic and metal particles (Lochner et al., 2011). A mouse osteolysis model has been used in four studies to investigate VEGF anditsreceptorsinasepticloosening.UHMWPEparticlesinducedinflammatoryosteoclastogenesis by up-regulating the expression of VEGF and its receptors 1 and 2 as well as RANK and RANKL (Ren et al., 2006). VEGF inhibitor reduced tissue inflammation(cellularinfiltration,membraneproliferation,andexpressionofIL-1βandTNF-α)aswellasboneresorption(Renetal.,2007).Antiangiogenicandantiosteolytic erythromycin treatment had the same effects but also reduced VEGF and VEGFR1 expression (Markel et al., 2009). Again in the fourth study, exposure toUHMWPEparticlesinducedinflammatoryosteolysis,whichwasassociatedwithincreased expression of VEGF and VEGFR1. Treatment with a VEGF neutralizing antibodysignificantlydiminishedthe inflammatoryosteolysis,andreducedtheexpression of VEGF and VEGFR1. However, treatment with a VEGFR2 inhibitor showednoeffectoninflammatoryosteolysis(Renetal.,2011).Inarecenttitaniumparticle inducedtissue inflammationmouseairpouchmodelstudy localVEGFinhibition(viaRNAinhibition)resultedinreducedexpressionofVEGF,TNF-α,IL-1βandRANKL.Thus,theresearcherssuggestedthatVEGFinhibitionmightbeapromisingtherapeuticcandidatetoalleviateparticle-inducedinflammation(Zhang et al., 2011).

Neovascularisation is essential in wound healing, tissue repair and remodelling and plays an important role in chronic inflammatory conditions like asepticloosening. VEGF is a powerful proangiogenic growth factor and induces proliferation, permeability, invasion, migration, survival and activation of endothelial cells (Parikh et al., 2004). VEGF coordinates endothelial cell invasion and extracellular matrix degradation by induction of MMPs, uPA, urokinase-type plasminogen activator receptor and tissue-type plasminogen activator all of which have been shown to be up-regulated around loose THRs (Parikh et al., 2004; Takagi et al., 1994; Nordsletten et al., 1996). VEGF modulates endothelial cell activation in part by up-regulating integrin expression, which is up-regulated around loose THRs (Parikh et al., 2004;

71

Li et al., 2000). VEGF can affect EC migration by activating focal adhesion kinase and the p38 pathway (Parikh et al., 2004). VEGF-induced NO synthesis contributes both to EC migration and permeability (Neufeld et al., 1999; Parikh et al., 2004). There is strong evidence that NO plays a major role in aseptic loosening (Hukkanen et al., 1997). VEGF seems to have an important role in modulating angiogenic effects of RANKL under physiological and pathological conditions mainly through VEGFR2 pathway (Min et al., 2003). RANKL has already been found to be important in aseptic loosening and osteolysis (Mandelin et al., 2003).

Based on literature and results of the study VI, it is very likely that the VEGF/VEGFRsystemisup-regulatedatleastbysomepro-inflammatorycytokinesandisessential in pathological angiogenesis in aseptic loosening. VEGF-based angiogenesis is important also in the development, remodelling and maintenance of PTs surrounding the loosening implant. The VEGF/VEGFR system is clearly osteogenic and probably important in coupling bone formation with angiogenesis. It is fairly safe to assume that it assists in bone formation in periprosthetic tissues of aseptically loosened hip implants in an attempt to keep up with the enhanced bone resorption. At the same it also may assist in local osteolysis during the aseptic loosening process in different PT sites via its known effects on preosteoclasts and OCs.

patient and control sample selection

Synovial joint samples from osteoarthritic joints or traumatized joints have been used in almost all immunohistochemical studies of aseptic loosening as control samples and were therefore chosen. In study III, we used the thin connective tissuesamplessurroundingwellfixedmetalplatesto imitateconnectivetissuesurroundingofwell-fixedhipimplants.Unfortunately,suchsamplesareusuallytiny and hard to get. Probably the best control material for rTHR samples would be hippseudocapsulesamplessurroundingwellfixedtotalhipreplacementimplants.Suchsamplesareforobviousreasonsverydifficulttoget,butcouldbeobtainedfrom revision operations performed for implant or peri-prosthetic bone fracture. Aseptic loosening, although developing slowly, is a dynamic process and not likely to proceed uniformly all around the implant. Due to this topographical heterogeneity, variouscharacteristicsofinflammationprobablyvarydependingonthesitefromwhich the sample was collected. In these studies samples were collected from lesions, which included synovial-like lining. Histologically they were found to be characterized by a rather uniform chronic foreign body reaction. Cemented and uncementedfixationanddifferent implantmodelswereused inpatients.Thenumberofpatientswastoosmalltodiscoverdifferencesbetweendifferentfixationmethods and implant models. Statistical testing was used to compare patient

72

and control samples, but multiple testing may have produced some false positive findings.Furthermore,statisticalsignificanceisnotnecessarilysynonymouswithclinicalsignificance.

releVance oF methods

The basic method used in all publications was indirect immunohistochemistry. It is more sensitive than the direct staining method because multiple secondary chromogen labelled antibodies are able to attach to primary antibodies, which results in improved intensive contrast. Background staining and false positive/negative staining can be problematic in immunohistochemistry. We used various methods to overcome these problems. Optimal concentration of antibodies was determinedinordertodiluteoutserumproteinsgivingnon-specificbackgroundstaining.Thisispossibleduetothehighaffinityofspecificantibodiesagainsttheirantigenic epitopes. Endogenous peroxidase in tissue sections was blocked with a relatively high concentration of hydrogen peroxide in methanol: locally produced reactive oxygen species are then able to denature the active site of the horse radish peroxidase although it is a rather stable enzyme (and therefore much used in immunohistochemistry). We used normal non-immune serum from the same speciestosaturatethesiteswhichmightnon-specificallybindsecondaryantibodies.As a negative staining control we used the same concentration of normal IgG of thesamesubtypebutwithirrelevantspecificityinsteadoftheprimaryantibody.Therefore,thedifferenceinimmunostainingbetweenthespecificallystainedsampleandthenegativestainingcontrolshouldbeduetotheantigen-specificprimaryantibodies, because all the other reagents and steps in the staining protocol are otherwise the same.

Theinter-andintra-samplevariationinhistopathologyandincytokineprofileof IT is quite high (Goodman et al., 1996). The heterogeneity is due to different mechanical and biologic environments along the bone-implant interface caused by/ leading to variable anatomy, vascularity and levels of wear debris in different areas. Cells in different areas are undergoing different rates of replication and metabolic activity. Revision THR samples were harvested from macroscopically active looking sites covered with synovial-like lining. In publications I-IV we used randommanualselectionofrepresentativemicroscopefieldsforimageanalysis,the same cut-off value for the staining intensity between different slides (between positive and negative) and semi-automatic cell counting was used. For many immunohistochemical markers the difference between the patient and control samples was relatively large so it was possible to use a relatively crude ad hoc grading system in publications V and VI, which may have diminished the sensitivity

73

of the analysis, but is less likely to lead to false results and interpretations due to the topographical variation in tissue histology.

Many of the cells in synovial membrane-like IT and control synovial membrane canbeidentifiedbytheirlocationandorganizationapartfromthemorphologyofindividualcells.InpublicationsVandVIweusedtheterm“stromalfibroblast-likecells” to describe spindle shaped connective tissue cells mostly organised in the same orientationalongthecollagenfibres.Mostofthesecellsarefibroblastsaccordingtoourpreviousfindingsbasedontheuseoffibroblastmarkersinimmunostaining(Santavirta et al., 1990a; 1990b; 1998). However, some of these cells are spindle-shaped monocyte/macrophages. The term “interstitial monocyte/macrophage-like cells” was used to describe mononuclear, non-spindle shaped, rounded interstitial cellsoften forming inflammatory cells infiltrates andcontainingphagocytosedimplant-derived debris. Majority of these are monocyte/macrophages according toourpreviousfindingswithmonocyte/macrophagespecificmarkers(Santavirtaet al., 1990a; 1990b; 1998).

We did not study bone cells and bone samples in this thesis work and it must be remembered that cellular events taking place at the bone surface are naturally critical for osseointegration and peri-implant osteolysis. Furthermore, the rTHR samples were taken during revision surgery when the implants have usually been loose already for a longer time. Mechanical forces in the IT area differ markedly between late and early phase of aseptic loosening when osteolysis is evident, but theimplantisnotyetloose.ThecytokineprofilemightbedifferentatearlierstagesbecausethelatestagerTHRsamplesreflecttheendstageofasepticloosening.

possiBle roles oF the studied growth Factors in aseptic loosening

All of the growth factors studied in this thesis work have multiple and partly overlappingeffectsonthetypesofcellsandtissuessurroundingwellfixedandasepticallylooseninghipreplacementimplantcomponentsandtheartificialjointitself as discussed in the literature review. On the other hand, the effects of the growth factors on individual cell types can be opposite depending on local tissue conditions and systemic and local regulatory mechanisms, i.e. on the context of the stimulation. Most of the studied growth factors are able to stimulate and/or inhibit the effects of each other on different cell types in different in vitro experiments. Furthermore, many of the locally up-regulated cytokines around THRs are able to stimulate and/or inhibit the studied growth factors. The relationships among phenotypically variable IT cells, ECM and ECM receptors, cytokines and their receptors, proteases and intracellular signalling pathways during development and

74

maintenance of the IT in aseptic loosening represent highly complex events that are regulated by a myriad of macro- and microenvironmental factors in an apparently perpetualstateofdynamicinterdependentflux.Itisachallengeforvarioushighthroughput methods (omics) and bioinformatics to produce a holistic view of the looseningandofthevariouspathwaysmediatingit.Itisverydifficulttoidentifythekey molecules and biosignatures in such complex biologic systems. Therefore, it is practically impossible to explain the exact mechanism of the studied growth factors in the aseptic loosening process. Merely, we were able to show that they all play somekindofrolesinthelooseningprocess.Morespecifically,wesuggestthattheroles are not only in bone formation but also in some extent in bone resorption, in IT development, remodelling and maintenance and evidently also in angiogenesis which markedly affects all these phenomena in aseptic loosening.

Our original hypothesis was that bone resorption is not normally coupled to bone formation in aseptic loosening. At the initiation of the study, it had already been reported that many cytokines and differentiation factors, which promote osteolysis are increased in PTs. It could have been possible that the anabolic factors stimulating new bone formation would have similarly been pathologically changed, being all low. This would naturally contribute to net loss of bone. The results of this thesis work seem to refute such an idea: growth factor containing cells enhancing bone formation were in general more frequent in PTs than in control tissues, with the exception of IGFs (IV). This suggests that coupling takes place also in the pathological environment of aseptic loosening and that both bone resorption and bone formation are enhanced but deranged in peri-implant bone. This assumption hasbeenconfirmedbydynamichistomorphometricstudiesofperi-implantbone(Takagi et al., 2001).

The observed down-regulation of IGFs in aseptic loosening (IV) could be explained with some kind of negative feedback regulation by other cytokines. For examplebFGF,TGF-βandPDGFcaninhibitIGF-Iand-IIsynthesisinosteoblasts(Conover and Rosen, 2002).

Bone turnover was shown to be up-regulated at least three years post operatively aroundwell-fixedTHRs(Venesmaaetal.,2012).Inasepticloosening,osteoblasticbone formation seems also to be up-regulated and this accelerated bone remodelling results in the formation of poor-quality bone favouring osteolysis (Takagi et al., 2001). The high periprosthetic bone turnover is in line with the increased production of both osteolysis and bone formation promoting cytokines in PTs. Most studies on the non-surgical treatment of aseptic loosening have focused on diminishing inflammationandosteolysis.Itmightbereasonabletostimulateboneformationat the same time.

This study expanded our knowledge of the role of anabolic growth factors in the aseptic loosening process. Despite evolving, sophisticated laboratory research

75

methods and slowly revealing complex cytokine interactions it seems that inhibiting aseptic loosening by manipulating cytokines is very challenging.

Theeffectsofpro-inflammatory/osteolyticcytokines,whichdown-regulateosteoprogenitor cell proliferation, differentiation and maturation could perhaps be inhibited by a combination of cytokine modulators or perhaps sequential use of such drugs. Extensive in vitro studies with mixed cell types such as monocyte/macrophages,fibroblasts,endothelialcells,osteoprogenitorcellsandOCsexposedto different wear particles and different cytokine cocktails could help to resolve this issue. Some mixture of different osteogenetic growth factors together with antibodies againstpro-inflammatorycytokinesmightchangethebalancebetweenosteolysisand bone formation in the favour of formation, which should be also studied in experimental animal models displaying accelerated peri-implant bone loss.

A special problem in therapeutic use of growth factors is their short biological half-life. Many strategies, such as use of polymers, pumps and coated implants have been investigated as possible methods for achieving constant growth factor levels at a given site in skeletal tissues with limited success. Gene therapy might be helpful in this respect in aseptic loosening (Ulrich-Vinther, 2007) perhaps combined with osteoconductive biomaterials like bioglass (Välimäki et al., 2005). It must berememberedthatasepticlooseningmightbeavoidedinthefutureifaspecificfixationmethodorimplantdesignresistingasepticlooseningcouldbedeveloped.

76

6 CONCLUSIONS

based on the results of the present studies it can be concluded that:

1. Itwasconfirmedthathistopathologicalfeaturesofinterfacetissue(IT)areverysimilar to synovial tissue from OA, except for the presence of wear particles and a chronic foreign body reaction in the former. These tissues consisted mainly ofvascularisedconnectivetissuecontaininghistiocytesandfibroblastsandcovered by synovial lining. IT from aseptic loosening contains particulate debris suchasUHMWPE,PMMA,andmetalwearandinflammatorymacrophages.These results indicate that particulate debris wear induces a biological reaction necessary in aseptic loosening of THR.

2. This study demonstrated the presence, cellular localization and extent of expressionofbFGF,TGF-β1and-β2,PDGF,IGF-Iand-II,BMP-2,-4, -6,and -7 and VEGF and VEGF receptors 1 and 2 in IT around loose THRs. The expression of these cytokines suggests that coupling of bone resorption to bone formation takes place also in the pathological environment of aseptic loosening of THR.

3. PresenceofbFGF,TGF-β1and-β2,PDGF,IGF-Iand-IIinpseudocapsulartissue obtained from patients undergoing revision THR was also found. These results indicate that pseudocapsular tissues also contribute to aseptic loosening of THR.

4. The number of cytokine positive cells was relatively similar in tissues from prostheseswithdifferent typesoffixation(cementandcementless)and/oralloy (CoCrMo and TiAlV). It seems that once a cellular host reaction occurs in the periprosthetic tissues and causes loosening, it appears to induce cytokines regardlessoftypeoffixationorprostheticmaterials.

5. The hypothesis of uniformly diminished production of cytokines able to stimulate bone formation as an important factor in aseptic loosening was proven tobewrong:suchacytokineprofilewasnotfoundinPTs.

6. Aseptic loosening is so multifactorial and complex process that it can not be explained solely by cytokine imbalances.

77

7 ACKNOWLEDGEMENTS

This work was carried out at the Department of Anatomy at the University of Helsinki, the ORTON Research Institute and the Orthopaedic Hospital of the Invalid Foundation, and Department of Orthopaedics and Traumatology, and Department of Medicine/ Invärtes medicin, Helsinki University Central Hospital, Finland during the years 1995-2013.

I wish to express my deepest gratitude to my supervisor Professor Yrjö T. Konttinen, MD, PhD, Head of TULES Research Group, Biomedicum Helsinki and Invalid Foundation. Without his encouraging attitude, his continued support and broad understanding in sciences, his constructive and timely discussions, and his creation of a free yet stimulating academic atmosphere, I would never have succeeded in accomplishing this work.

My sincere thanks go to my late supervisor Professor Seppo Santavirta, MD, PhD, Department of Orthopaedics and Traumatology, Helsinki University Central Hospital, for providing me the opportunity to work at his department. His famous work in orthopaedics, his never failing support, persistent encouragement and valuable criticism have inspired to the accomplishment of this work.

I would also like to thank my supervisor Docent Dan Nordström, MD, PhD, Department of Medicine Division of Internal Medicine, Helsinki University Central Hospital, for teaching and advising me in the use of immunostaining methods during the early phases of this work and for his support and valuable comments andsuggestionsduringthefinalphasesofit.

I am most grateful to late Professor Ismo Virtanen, Head of the Department of Anatomy, University of Helsinki; Professor Emeritus Pär Slätis, and the late Professor Antti Alho, both former Heads of the Invalid Foundation, for all placing the research facilities of their departments at my disposal.

I am indebted to the official pre-examiners, docent Petri Virolainen anddocent Nils Hailer. After they reviewed the manuscript of this thesis, it improved significantly.Theyarealsoacknowledgedforpositivecriticismandencouragement.

I am deeply grateful to all of my co-authors from Finland, Norway, USA and Italy.I wish to thank all my senior and junior colleagues of the TULES Research Group

for the pleasant and collaborative atmosphere. I wish to express my great thanks to Dr. Jing-Wen Xu, Dr. Nina Santavirta, PhD Svetlana Solovieva, Dr. Tian-Fang Li,Dr.JianMa,Dr.OliverMichelsson,Dr.JanLassus,Dr.ArnoldasČeponis,andDr. Sirajedin Natah for their support and friendship. Especially, would like to give my warmest thanks to Eija Kaila, Mari Ainola, Mika Hukkanen, Jami Mandelin, Mikko Liljeström, Pauliina Porola and Erkki Hänninen for their skilful help and

78

friendship. I also wish to thank other members not mentioned here in the TULES Research Group.

Special thanks go to my cousin, good friend and TULES Group colleague Eero Waris, who has helped me in many ways during all these years. I am indebted to all doctors at the Department of Orthopaedics and Traumatology, Helsinki University Central Hospital, for providing the samples used in this study.

I wish to express my greatest thanks to my family for their continuous and alwaysdeeplove, infinitivesupportandunderstandingduringall theseyears inHelsinki, Kuopio and Mikkeli.

ThisworkwasfinanciallysupportedbytheFinnishAcademy,theNationalCentreof Excellence in Biomaterials of the Academy of Finland, Sigrid Jusélius Foundation, University of Helsinki, Helsinki University Central Hospital, Kuopio University Hospital, Mikkeli Central Hospital, Finska Läkaresällskapet, The National Graduate School for Clinical Investigation of the Finnish Academy, The Finnish Orthopaedic andTraumatologySociety,OscarÖflundFoundation,PehrOscarKlingendahlFoundation, Danish Council for Strategic Research, The Finnish Medical Foundation and Finnish Cultural Foundation/ South Savo Regional Fund.

79

REFERENCES

Abramson SB, Attur M. Developments in the scientific understanding of osteoarthritis. Arthritis Res Ther. 2009;11(3):227.

Agarwal R, Williams K, Umscheid CA, Welch WC. Osteoinductive bone graft substitutes for lumbar fusion: a systematic review. J Neurosurg Spine. 2009 Dec;11(6):729-40.

Akeno N, Robins J, Zhang M, Czyzyk-Krzeska MF, Clemens TL. Induction of vascular endothelial growth factor by IGF-I in osteoblast-like cells is mediated by the PI3K signaling pathway through the hypoxia-inducible factor-2alpha. Endocrinology. 2002 Feb;143(2):420-5.

Alexander KA, Chang MK, Maylin ER, et al. Osteal macrophages promote in vivo intramembranous bone healing in a mouse tibial injury model. J Bone Miner Res. 2011 Jul;26(7):1517-32.

Aldridge SE, Lennard TW, Williams JR, Birch MA. Vascular endothelial growth factor acts as an osteolytic factor in breast cancer metastases to bone. Br J Cancer. 2005 Apr 25;92(8):1531-7.

Allori AC, Sailon AM, Warren SM. Biological basis of bone formation, remodeling, and repair. Tissue Eng Part B Rev. 2008;14(3):259-93.

al-Saffar N. The osteogenic properties of the interface membrane at the site of orthopedic implants: the impact of underlying joint disease. J Long Term Eff Med Implants. 1999;9(1-2):23-45.

Andersson MK, Anissian L, Stark A, Bucht E, Felländer-Tsai L, Tsai JA. Synovial fluidfromloosehiparthroplastiesinhibitshumanosteoblasts.ClinOrthopRelat Res. 2000 Sep;(378):148-54.

AnderssonMK,StarkA,AnissianL,MohanS,TsaiJA.LowIGF-Iinsynovialfluidand serum in patients with aseptic prosthesis loosening. Acta Orthop. 2005 Jun;76(3):320-5.

Andrae J, Gallini R, Betsholtz C. Role of platelet-derived growth factors in physiology and medicine. Genes Dev. 2008 May 15;22(10):1276-312.

Appel AM, Sowder WG, Siverhus SW, Hopson CN, Herman JH. Prosthesis-associated pseudomembrane-induced bone resorption. Br J Rheumatol 1990;29:32-6.

Armelin HA. Pituitary extracts and steroid hormones in the control of 3T3 cell growth. Proc Natl Acad Sci U S A. 1973 Sep;70(9):2702-6.

80

Aro HT, Govender S, Patel AD, et al. Recombinant human bone morphogenetic protein-2: a randomized trial in open tibial fractures treated with reamed nailfixation.JBoneJointSurgAm.2011May4;93(9):801-8.

AthanasouNA,QuinnJ,BulstrodeCJ.Resorptionofbonebyinflammatorycellsderived from the joint capsule of hip arthroplasties. J Bone Joint Surg Br. 1992 Jan;74(1):57-62.

Avruch J, Khokhlatchev A, Kyriakis JM, Luo Z, Tzivion G, Vavvas D, Zhang XF. Ras activation of the Raf kinase: tyrosine kinase recruitment of the MAP kinase cascade. Recent Prog Horm Res. 2001;56:127-55.

Baron RE. Anatomy and ultrastructure of bone. In: Favus MJ (eds) Primer on the metabolic bone diseases and disorders of mineral metabolim. Lippincott-Raven, New York, 1996:3-11.

Barrientos S, Stojadinovic O, Golinko MS, Brem H, Tomic-Canic M. Growth factors and cytokines in wound healing. Wound Repair Regen. 2008 Sep-Oct;16(5):585-601.

Beamer B, Hettrich C, Lane J. Vascular endothelial growth factor: an essential component of angiogenesis and fracture healing. HSS J. 2010 Feb;6(1):85-94.

Beck RT, Illingworth KD, Saleh KJ. Review of periprosthetic osteolysis in total joint arthroplasty: An emphasis on host factors and future directions. J Orthop Res. 2011 Sep 15. doi: 10.1002/jor.21554.

Behr B, Leucht P, Longaker MT, Quarto N. Fgf-9 is required for angiogenesis and osteogenesis in long bone repair. Proc Natl Acad Sci U S A. 2010 Jun 29;107(26):11853-8.

Berse B, Hunt JA, Diegel RJ, Morganelli P, Yeo K, Brown F, Fava RA. Hypoxia augments cytokine (transforming growth factor-beta (TGF-beta) and IL-1)-induced vascular endothelial growth factor secretion by human synovial fibroblasts.ClinExpImmunol.1999Jan;115(1):176-82.

Betts F, Wright T, Salvati EA, Boskey A, Bansal M. Cobalt-alloy metal debris in periarticular tissues from total hip revision arthroplasties. Metal contents andassociatedhistologicfindings.ClinOrthop1992;276:76-82.

Bloebaum RD, Zou L, Bachus KN, Shea KG, Hofmann AA, Dunn HK. Analysis of particles in acetabular components from patients with osteolysis. Clin Orthop 1996;338:109-18.

Bobacz K, Ullrich R, Amoyo L, Erlacher L, Smolen JS, Graninger WB. Stimulatory effects of distinct members of the bone morphogenetic protein family on ligamentfibroblasts.AnnRheumDis.2006Feb;65(2):169-77.

BonewaldLF.Transforminggrowthfactor-β.In:Principlesofbonebiology(eds.Bilezikian JP, Raisz LG, Rodan GA). Academic Press, San Diego, CA, U.S.A. 1996: 647- 59.

Bonewald LF. The amazing osteocyte. J Bone Miner Res. 2011 Feb;26(2):229-38.

81

Bramlage CP, Häupl T, Kaps C, et al. Decrease in expression of bone morphogenetic proteins 4 and 5 in synovial tissue of patients with osteoarthritis and rheumatoid arthritis. Arthritis Res Ther. 2006;8(3):R58.

Brown TD, Lundberg HJ, Pedersen DR, Callaghan JJ. 2009 Nicolas Andry Award: clinical biomechanics of third body acceleration of total hip wear. Clin Orthop Relat Res. 2009;467(7):1885-97.

Buly RL, Huo HM, Salvati E, Brien W, Bansal M. Titanium wear debris in failed cemented total hip arthroplasty: An analysis of 71 cases. J Arthroplasty 1992;7:315-23.

Burr DB, Gallant MA. Bone remodelling in osteoarthritis. Nat Rev Rheumatol. 2012 Nov;8(11):665-73.

Caetano-Lopes J. Effect of inflammation on bone biological, structural andmechanical behaviour. Thesis, University of Lissabon, 2011.

Campbell PA, Wang M, Amstutz HC, Goodman SB. Positive cytokine production in failed metal-on-metal total hip replacements. Acta Orthop Scand. 2002 Oct;73(5):506-12.

Canalis E, Pash J, Gabbitas B, Rydziel S, Varghese S. Growth factors regulate the synthesis of insulin-like growth factor-I in bone cell cultures. Endocrinology. 1993 Jul;133(1):33-8.

Canalis E, Economides AN, Gazzerro E. Bone morphogenetic proteins, their antagonists, and the skeleton. Endocr Rev. 2003 Apr;24(2):218-35.

Canalis E. Growth factor control of bone mass. J Cell Biochem. 2009 Nov 1;108(4):769-77.

Caplan AI, Correa D. PDGF in bone formation and regeneration: new insights into a novel mechanism involving MSCs. J Orthop Res. 2011 Dec;29(12):1795-803.

Carlsson AS, Gentz CF, Linder L. Localized bone resorption in the femur in mechanical failure of cemented total hip arthroplasties. Acta Orthop Scand 1983;54:396-402.

Castillo EF, Schluns KS. Regulating the immune system via IL-15 transpresentation. Cytokine. 2012 Sep;59(3):479-90.

Cenni E, Savarino L, Baldini N, Rotini R, Marinelli A, Mieti M, Giunti A. Plasma levelsofcoagulation inhibitors,fibrinolyticmarkersandplatelet-derivedgrowth factor-AB in patients with failed hip prosthesis. Acta Orthop Scand. 2003 Oct;74(5):559-64.

Cenni E, Savarino L, Antonioli D, Pellacani A, Giunti A, Baldini N. Plasma levels of platelet-derived growth factor BB and transforming growth factor in patients with failed hip prostheses. Acta Orthop. 2005 Feb;76(1):61-6.

Chang J, Most D, Thunder R, Mehrara B, Longaker MT, Lineaweaver WC. Molecular studiesinflexortendonwoundhealing:theroleofbasicfibroblastgrowthfactor gene expression. J Hand Surg Am. 1998 Nov;23(6):1052-8.

82

Charnley J. Arthroplasty of the hip: a new operation. Lancet 1961;1:1129-1132.Charnley J. The bonding of prostheses to bone by cement. J Bone Joint Surg

1964;46B:518-29.Charnley J, Follacci FM, Hammond BT. The long-term reaction of bone to self-

curing acrylic cement. J Bone Joint Surg 1968;50B:822-9.Charnley J. Low friction arthroplasty of the hip. New York, Springer-Verlag, 1979.ChenG,DengC,LiYP.TGF-βandBMPsignalinginosteoblastdifferentiationand

bone formation. Int J Biol Sci. 2012;8(2):272-88. Ching YH, Sutton TL, Pierpont YN, Robson MC, Payne WG. The use of growth

factors and other humoral agents to accelerate and enhance burn wound healing. Eplasty. 2011;11:e41.

ChintalgattuV,NairDM,KatwaLC.Cardiacmyofibroblasts:anovelsourceofvascular endothelial growth factor (VEGF) and its receptors Flt-1 and KDR. J Mol Cell Cardiol. 2003 Mar;35(3):277-86.

Chiba J, Rubash HE, Kim KJ, Iwaki Y. The characterisation of cytokines in the interface tissue obtained from failed cementless total hip arthroplasty with and without femoral osteolysis. Clin Orthop 1994;300:304-12.

Clarke B. Normal bone anatomy and physiology. Clin J Am Soc Nephrol. 2008 Nov;3 Suppl 3:S131-9.

Clarkin CE, Emery RJ, Pitsillides AA, Wheeler-Jones CP. Evaluation of VEGF-mediated signaling in primary human cells reveals a paracrine action for VEGF in osteoblast-mediated crosstalk to endothelial cells. J Cell Physiol. 2008;214(2):537-44.

Cohick WS, Clemmons DR. The insulin-like growth factors. Annu Rev Physiol. 1993;55:131-53.

Collin-OsdobyP,RotheL,BekkerS,AndersonF,HuangY,OsdobyP.Basicfibroblastgrowth factor stimulates osteoclast recruitment, development, and bone pit resorption in association with angiogenesis in vivo on the chick chorioallantoic membrane and activates isolated avian osteoclast resorption in vitro. J Bone Miner Res. 2002 Oct;17(10):1859-71.

Conover CA, Rosen C. The role of insulin-like growth factors and binding proteins in bone cell biology. In Bilezikian JP, Raisz LG, Rodan Ga eds. Principles of bone biology, 2ed. San Diego: Academic Press 2002:225-73.

Cook PW, Coffey RJ Jr, Magun BE, Pittelkow MR, Shipley GD. Expression and regulationofmRNAcodingforacidicandbasicfibroblastgrowthfactorandtransforming growth factor alpha in cells derived from human skin. Mol Endocrinol. 1990 Sep;4(9):1377-85.

Cook SD, Barrack RL, Clemow AJT. Corrosion and wear at the modular interface of uncemented femoral stems. J Bone Joint Surg 1994;76B:68-72.

83

CoolSM,SnymanCP,NurcombeV,ForwoodM.Temporalexpressionoffibroblastgrowth factor receptors during primary ligament repair. Knee Surg Sports Traumatol Arthrosc. 2004 Sep;12(5):490-6.

Cui Y, Jean F, Thomas G, Christian JL. BMP-4 is proteolytically activated by furin and/or PC6 during vertebrate embryonic development. EMBO J. 1998 Aug 17;17(16):4735-43.

Dai J, Rabie AB. VEGF: an essential mediator of both angiogenesis and endochondral ossification.JDentRes.2007Oct;86(10):937-50.

Darland DC, Massingham LJ, Smith SR, Piek E, Saint-Geniez M, D'Amore PA. Pericyte production of cell-associated VEGF is differentiation-dependent and is associated with endothelial survival. Dev Biol. 2003;264(1):275-88.

Dean DD, Schwartz Z, Blanchard CR, et al. Ultrahigh molecular weight polyethylene particles have direct effects on proliferation, differentiation, and local factor production of MG63 osteoblast-like cells. J Orthop Res. 1999 Jan;17(1):9-17.

Deckers MM, Karperien M, van der Bent C, et al. Expression of vascular endothelial growth factors and their receptors during osteoblast differentiation. Endocrinology. 2000 May;141(5):1667-74.

Deckers MM, van Bezooijen RL, van der Horst G, et al. Bone morphogenetic proteins stimulate angiogenesis through osteoblast-derived vascular endothelial growth factor A. Endocrinology. 2002 Apr;143(4):1545-53.

Devescovi V, Leonardi E, Ciapetti G, Cenni E. Growth factors in bone repair. Chir Organi Mov. 2008 Dec;92(3):161-8.

Dorr LD, Bloebaum R, Emmanual J, Meldrum R. Histologic, biochemical and ion analysisoftissueandfluidsretrievedduringtotalhiparthoplasty.ClinOrthop1990;261:82-95.

Dwek JR. The periosteum: what is it, where is it, and what mimics it in its absence? Skeletal Radiol. 2010 Apr;39(4):319-23.

Eriksen EF. Cellular mechanisms of bone remodeling. Rev Endocr Metab Disord. 2010;11(4):219-27.

Erlebacher A, Derynck R. Increased expression of TGF-beta 2 in osteoblasts results in an osteoporosis-like phenotype. J Cell Biol. 1996 Jan;132(1-2):195-210.

Eskelinen A, Paavolainen P, Helenius I, Pulkkinen P, Remes V. Total hip arthroplasty for rheumatoid arthritis in younger patients: 2,557 replacements in the Finnish Arthroplasty Register followed for 0-24 years. Acta Orthop. 2006 Dec;77(6):853-65.

Evans EM, Freeman MAR, Miller AJ, Vernon-Roberts B. Metal sensitivity as a cause of bone necrosis and loosening of the prosthesis in total joint replacement. J Bone Joint Surg 1974;56B:626-42.

84

Felson DT, Chaisson CE, Hill CL, et al. The association of bone marrow lesions with pain in knee osteoarthritis. Ann Intern Med. 2001 Apr 3;134(7):541-9.

Ferrara N, Gerber HP, LeCouter J. The biology of VEGF and its receptors. Nat Med. 2003 Jun;9(6):669-76.

Fiedler J, Etzel N, Brenner RE. To go or not to go: Migration of human mesenchymal progenitor cells stimulated by isoforms of PDGF. J Cell Biochem. 2004 Nov 15;93(5):990-8.

Fiorito S, Magrini L, Goalard C. Pro-inflammatory and anti-inflammatory circulating cytokines and periprosthetic osteolysis. J Bone Joint Surg Br. 2003 Nov;85(8):1202-6.

FredrikssonL,LiH,ErikssonU.ThePDGFfamily:fourgeneproductsformfivedimeric isoforms. Cytokine Growth Factor Rev. 2004 Aug;15(4):197-204.

Gabbitas B, Pash J, Canalis E. Regulation of insulin-like growth factor-II synthesis in bone cell cultures by skeletal growth factors. Endocrinology. 1994 Jul;135(1):284-9.

Gallo J, Goodman SB, Konttinen YT, Raska M. Particle disease: Biologic mechanisms of periprosthetic osteolysis in total hip arthroplasty. Innate Immun. 2013;19(2):213-24.

Garimella R, Tague SE, Zhang J, et al. Expression and synthesis of bone morphogenetic proteins by osteoclasts: A possible path to anabolic bone remodeling. J Histochem Cytochem. 2008;56:569-577.

Geffner ME, Bersch N, Lippe BM, Rosenfeld RG, Hintz RL, Golde DW. Growth hormone mediates the growth of T-lymphoblast cell lines via locally generated insulin-like growth factor-I. J Clin Endocrinol Metab. 1990 Aug;71(2):464-9.

Giatromanolaki A, Sivridis E, Maltezos E, et al. Upregulated hypoxia inducible factor-1alpha and -2alpha pathway in rheumatoid arthritis and osteoarthritis. Arthritis Res Ther. 2003;5(4):R193-201.

Giatromanolaki A, Sivridis E, Athanassou N, et al. The angiogenic pathway "vascular endothelialgrowthfactor/flk-1(KDR)-receptor"inrheumatoidarthritisandosteoarthritis. J Pathol. 2001 May;194(1):101-8.

Giustina A, Mazziotti G, Canalis E. Growth hormone, insulin-like growth factors, and the skeleton. Endocr Rev. 2008 Aug;29(5):535-59.

Goel S, Duda DG, Xu L, Munn LL, Boucher Y, Fukumura D, Jain RK. Normalization of the vasculature for treatment of cancer and other diseases. Physiol Rev. 2011 Jul;91(3):1071-121.

Goldring SR, Schiller AL, Roelke M, Rourke CM, O’Neill DA, Harris WH. The synovial-like membrane at the bone-cement interface in loose total hip replacements and its proposed role in bone lysis. J Bone Joint Surg 1983;63A:575-84.

85

Goldring SR, Jasty M, Roelke MS, Rourke CM, Bringhurst FR, Harris WH. Formation of a synovial-like membrane at the bone -cement interface: its role in bone resorption and implant loosening after total hip replacement. Arthritis Rheum 1986;29:836-42.

Goldring SR, Goldring MB. Cytokines and skeletal physiology. Clin Orthop 1996;324:13-23.

Goodman SB, Knoblich G, O'Connor M, Song Y, Huie P, Sibley R. Heterogeneity in cellularandcytokineprofilesfrommultiplesamplesoftissuesurroundingrevised hip prostheses. J Biomed Mater Res. 1996 Jul;31(3):421-8.

Goodman SB, Huie P, Song Y, Schurman D, Maloney W, Woolson S, Sibley R. Cellularprofileandcytokineproductionatprosthetic interfaces.Studyoftissues retrieved from revised hip and knee replacements. J Bone Joint Surg 1998;80B:531-9.

GospodarowiczD.Localisationofafibroblastgrowthfactoranditseffectaloneandwith hydrocortisone on 3T3 cell growth. Nature. 1974 May 10;249(453):123-7.

Govender S, Csimma C, Genant HK, Valentin-Opran A, et al.; BMP-2 Evaluation in Surgery for Tibial Trauma (BESTT) Study Group. Recombinant human bone morphogenetic protein-2 for treatment of open tibial fractures: a prospective, controlled,randomizedstudyoffourhundredandfiftypatients.JBoneJointSurg Am. 2002 Dec;84-A(12):2123-34.

Grant M, Jerdan J, Merimee TJ. Insulin-like growth factor-I modulates endothelial cell chemotaxis. J Clin Endocrinol Metab 1987;65:370-1.

Grant MB, Mames RN, Fitzgerald C, et al. Insulin-like growth factor I as an angiogenic agent. In vivo and in vitro studies. Ann NY Acad Sci 1993;692:230-42.

Hailer NP, Garellick G, Kärrholm J. Uncemented and cemented primary total hip arthroplasty in the Swedish Hip Arthroplasty Register. Acta Orthop. 2010 Feb;81(1):34-41.

Hallab NJ, Jacobs JJ. Biologic effects of implant debris. Bull NYU Hosp Jt Dis. 2009;67(2):182-8.

Happonen KE, Saxne T, Aspberg A, Mörgelin M, Heinegård D, Blom AM. Regulation of complement by cartilage oligomeric matrix protein allows for a novel molecular diagnostic principle in rheumatoid arthritis. Arthritis Rheum. 2010 Dec;62(12):3574-83.

Harris WH. The problem is osteolysis. Clin Orthop 1995;311:46-53.Hattersley G, Chambers TJ. Effects of transforming growth factor beta 1 on the

regulation of osteoclastic development and function. J Bone Miner Res. 1991 Feb;6(2):165-72.

Hayden JM, Mohan S, Baylink DJ. The insulin-like growth factor system and the coupling of formation to resorption. Bone 1995;17 Suppl:93-8.

86

Haynes DR, Crotti TN, Potter AE, Loric M, Atkins GJ, Howie DW, Findlay DM. The osteoclastogenic molecules RANKL and RANK are associated with periprosthetic osteolysis. J Bone Joint Surg Br. 2001 Aug;83(6):902-11.

Heinemann DE, Siggelkow H, Ponce LM, Viereck V, Wiese KG, Peters JH. Alkaline phosphatase expression during monocyte differentiation. Overlapping markers as a link between monocytic cells, dendritic cells, osteoclasts and osteoblasts. Immunobiology. 2000 May;202(1):68-81.

Heino TJ, Hentunen TA, Väänänen HK. Osteocytes inhibit osteoclastic bone resorption through transforming growth factor –beta: enhancement by estrogen. J Cell Biochem. 2002;85(1):185-97.

Hemingway F, Taylor R, Knowles HJ, Athanasou NA. RANKL-independent human osteoclast formation with APRIL, BAFF, NGF, IGF I and IGF II. Bone. 2011 Apr 1;48(4):938-44.

Hemmings BA, Restuccia DF. PI3K-PKB/Akt pathway. Cold Spring Harb Perspect Biol. 2012 Sep 1;4(9):a011189.

Hill PA, Tumber A, Meikle MC. Multiple extracellular signals promote osteoblast survival and apoptosis. Endocrinology. 1997 Sep;138(9):3849-58.

Hock JM, Canalis E. Platelet-derived growth factor enhances bone cell replication, but not differentiated function of osteoblasts. Endocrinology. 1994 Mar;134(3):1423-8.

Hollinger JO, Hart CE, Hirsch SN, Lynch S, Friedlaender GE. Recombinant human platelet-derived growth factor: biology and clinical applications. J Bone Joint Surg Am. 2008 Feb;90 Suppl 1:48-54.

Holt G, Murnaghan C, Reilly J, Meek RM. The biology of aseptic osteolysis. Clin Orthop Relat Res. 2007;460:240-52.

Howie DW, McGee MA, Pearcy MJ. The response to particulate debris. Orthop Clin North Am 1993;24:571-81.

HuiskesR,VerdonschotN,NivbrantB.Migration,stemshape,andsurfacefinishincemented total hip arthroplasty. Clin Orthop Relat Res. 1998 Oct;(355):103-12.

Hukkanen M, Corbett SA, Batten J, et al. Aseptic loosening of total hip replacement: macrophage expression of inducible nitric oxide synthase and cyclo-oxygenase-2, together with peroxynitrite formation, as a possible mechanism for early prosthesis failure. J Bone Joint Surg 1997;79B:467-74.

Huo MH, Salvati EA, Lieberman JR, Betts F, Bansal M. Metallic debris in femoral endosteolysis in failed cemented total hip arthroplasties. Clin Orthop 1992;276:157-68.

Huo MH, Dumont GD, Knight JR, Mont MA. What's new in total hip arthroplasty. J Bone Joint Surg Am. 2011 Oct 19;93(20):1944-50.

87

Ikeda M, Hosoda Y, Hirose S, Okada Y, Ikeda E. Expression of vascular endothelial growth factor isoforms and their receptors Flt-1, KDR, and neuropilin-1 in synovial tissues of rheumatoid arthritis. J Pathol. 2000 Aug;191(4):426-33.

Illgen RL 2nd, Forsythe TM, Pike JW, Laurent MP, Blanchard CR. Highly crosslinked vs conventional polyethylene particles--an in vitro comparison of biologic activities. J Arthroplasty. 2008 Aug;23(5):721-31.

Isaac C, Gharaibeh B, Witt M, Wright VJ, Huard J. Biologic approaches to enhance rotator cuff healing after injury. J Shoulder Elbow Surg. 2012 Feb;21(2): 181-90.

Ishiguro N, Kojima T, Ito T, et al. Macrophage activation and migration in interface tissue around loosening total hip arthroplasty components. J Biomed Mater Res. 1997 Jun 5;35(3):399-406.

Ishii T, Shimazu Y, Nishiyama I, Kikuta J, Ishii M. The role of sphingosine 1-phosphate in migration of osteoclast precursors; an application of intravital two-photon microscopy. Mol Cells. 2011 May;31(5):399-403.

Itoh H, Nasu K, Matsumoto H, Kawano Y, Yoshimatsu J, Narahara H. Hypoxia regulates vascular endothelial growth factor and soluble fms-like tyrosine kinase-1secretionbyhumanoviductalepithelialcellsandstromalfibroblasts.Fertil Steril. 2006 Apr;85 Suppl 1:1097-102.

Itonaga I, Sabokbar A, Murray DW, Athanasou NA. Effect of osteoprotegerin and osteoprotegerin ligand on osteoclast formation by arthroplasty membrane derived macrophages. Ann Rheum Dis. 2000 Jan;59(1):26-31.

Jasty M, Floyd WE, Schiller AL, Goldring SR, Harris WH. Localized osteolysis in stable, non-septic total hip replacement. J Bone Joint Surg 1986;68A:912-9.

Jacobs JJ, Hallab NJ. Loosening and osteolysis associated with metal-on-metal bearings: a local effect of metal hypersensitivity? J Bone Joint Surg Am 2006;88:1171–1172.

Jell GM, Al-Saffar N. Does a pro-angiogenic state exist in the bone-implant interface of aseptically loosened joint prosthesis? J Mater Sci Mater Med. 2001 Oct-Dec;12(10-12):1069-73.

JilkaRL,WeinsteinRS,BellidoT,ParfittAM,ManolagasSC.Osteoblastprogrammedcell death (apoptosis): modulation by growth factors and cytokines. J Bone Miner Res. 1998 May;13(5):793-802.

Jiranek WA, Machado M, Jasty M, et al. Production of cytokines around loosened cemented acetabular components. Analysis with immunohistochemical techniques and in situ hybridization. J Bone Joint Surg 1993;75A:863-79.

Johnsson A, Heldin CH, Westermark B, Wasteson A. Platelet-derived growth factor: identificationofconstituentpolypeptidechains.BiochemBiophysRes Commun. 1982 Jan 15;104(1):66-74.

Jones DA, Hungerford DS. Cement disease. Clin Orthop 1987;225:192-206.

88

Jämsen E, Peltola M, Eskelinen A, Lehto MU. Comorbid diseases as predictors of survival of primary total hip and knee replacements: a nationwide register-based study of 96 754 operations on patients with primary osteoarthritis. Ann Rheum Dis. 2013; 72:1975-1982.

Kadoya Y, Kobayashi A, Ohashi H. Wear and osteolysis in total joint replacements. Acta Orthop Scand Suppl. 1998 Feb;278:1-16.

Kaigler D, Avila G, Wisner-Lynch L, et al. Platelet-derived growth factor applications in periodontal and peri-implant bone regeneration. Expert Opin Biol Ther. 2011 Mar;11(3):375-85.

Kann S, Chiu R, Ma T, Goodman SB. OP-1 (BMP-7) stimulates osteoprogenitor cell differentiation in the presence of polymethylmethacrylate particles. J Biomed Mater Res A. 2010 Aug;94(2):485-8.

KelpkeSS,ZinnKR,RueLW,ThompsonJA.Site-specificdeliveryofacidicfibroblastgrowth factor stimulates angiogenic and osteogenic responses in vivo. J Biomed Mater Res A. 2004 Nov 1;71(2):316-25.

Kiaer T, Grønlund J, Sørensen KH. Subchondral pO2, pCO2, pressure, pH, and lactate in human osteoarthritis of the hip. Clin Orthop Relat Res. 1988 Apr;(229):149-55.

Kim KJ, Rubash HE, Wilson SC, D’Antonio JA, McClain EJ. A histologic and biochemical comparison of the interface tissues in cementless and cemented hip prostheses. Clin Orthop 1993;287:142-52.

Knowles HJ, Athanasou NA. Canonical and non-canonical pathways of osteoclast formation. Histol Histopathol. 2009 Mar;24(3):337-46.

Koizumi K, Saitoh Y, Minami T, et al. Role of CX3CL1/fractalkine in osteoclast differentiation and bone resorption. J Immunol. 2009 Dec 15;183(12):7825-31.

Konttinen YT, Imai S, Suda A. Neuropeptides and the puzzle of bone remodeling. State of the art. Acta Orthop Scand 1996a;67:632-9.

Konttinen YT, Kurvinen H, Takagi M, et al. Interleukin-1 and collagenases around loosening total hip prostheses. Clin Exp Rheumatol 1996b;14:255-62.

Konttinen YT, Xu JW, Patiala H, Imai S, Waris V, Li TF, Goodman SB, Nordsletten L, Santavirta S. Cytokines in aseptic loosening of total hip replacement. Curr Orthop 1997a;11:40-7.

Konttinen YT, Waris V, Xu JW, Jiranek WA, Sorsa T, Virtanen I, Santavirta S. Transforming growth factor-β 1 and 2 in the synovial-like interface membrane between implant and bone in loosening of total hip arthroplasty. J Rheumatol. 1997b Apr;24(4):694-701.

Konttinen YT, Zhao D, Beklen A, Ma G, Takagi M, Kivelä-Rajamäki M, Ashammakhi N, Santavirta S. The microenvironment around total hip replacement prostheses. Clin Orthop Relat Res. 2005 Jan;(430):28-38.

89

Konttinen YT, Tiainen VM, Gomez-Barrena E, Hukkanen M, Salo J. Innervation of the joint and role of neuropeptides. Ann N Y Acad Sci. 2006 Jun;1069:149-54.

Konttinen YT, Pajarinen J. Surgery: Adverse reactions to metal-on-metal implants. Nat Rev Rheumatol. 2012 Dec 4;9(1):5-6.

Konttinen YT, Sillat T, Barreto G, Ainola M, Nordström DCE. Osteoarthritis as an autoinflammatorydiseasecausedbychondrocyte-mediatedinflammatoryresponses. Arthritis Rheum 2012 Mar;64(3):613-6.

Koreny T, Tunyogi-Csapó M, Gál I, Vermes C, Jacobs JJ, Glant TT. The role of fibroblastsandfibroblast-derivedfactorsinperiprostheticosteolysis.ArthritisRheum. 2006 Oct;54(10):3221-32.

Kudo O, Sabokbar A, Pocock A, Itonaga I, Fujikawa Y, Athanasou NA. Interleukin-6 and interleukin-11 support human osteoclast formation by a RANKL-independent mechanism. Bone. 2003 Jan;32(1):1-7.

Kurtz SM, Roder C, Lau E, et al. International survey of primary and revision total hip replacement. Paper 365. Presented at the 56th Annual Meeting of the Orthopaedic Research Society. March 6-9, 2010. New Orleans

Laitinen M, Jortikka L, Halttunen T, et al. Measurement of total and local bone morphogenetic protein concentration in bone tumours. Int Orthop. 1997;21(3):188-93.

Lamberg A, Bechtold JE, Baas J, Søballe K, Elmengaard B. Effect of local TGF-beta1andIGF-1releaseonimplantfixation:comparisonwithhydroxyapatitecoating: a paired study in dogs. Acta Orthop. 2009 Aug;80(4):499-504.

Lambert C, Mathy-Hartert M, Dubuc JE, et al. Characterization of synovial angiogenesis in osteoarthritis patients and its modulation by chondroitin sulfate. Arthritis Res Ther. 2012 Mar 12;14(2):R58.

LennoxDW,SchofieldBH,McDonaldDF,RileyJrLH.Ahistologiccomparisonofasepticlooseningofcemented,press-fit,andbiologicingrowthprostheses.Clin Orthop 1987;225:171-91.

Li TF, Xu JW, Santavirta S, Nordsletten L, Michelsson O, Takagi M, Virtanen I, KonttinenYT.Distributionoffibronectinsandtheir integrinreceptors ininterface tissue from aseptic loosening of hip prostheses. Clin Exp Rheumatol. 2000 Mar-Apr;18(2):221-5.

Li J, Zhang YP, Kirsner RS. Angiogenesis in wound repair: angiogenic growth factors and the extracellular matrix. Microsc Res Tech. 2003 Jan 1;60(1):107-14.

LinJM,CallonKE,LinJS,etal.Actionsoffibroblastgrowthfactor-8inbonecellsin vitro. Am J Physiol Endocrinol Metab. 2009 Jul;297(1):E142-50.

Lindholm TC, Lindholm TS, Alitalo I, Urist MR. Bovine bone morphogenetic protein (bBMP) induced repair of skull trephine defects in sheep. Clin Orthop Relat Res. 1988 Feb;227:265-8.

90

Lissenberg-ThunnissenSN,deGorterDJ,SierCF,SchipperIB.Useandefficacyof bone morphogenetic proteins in fracture healing. Int Orthop. 2011 Sep;35(9):1271-80.

Longo UG, Lamberti A, Maffulli N, Denaro V. Tissue engineered biological augmentation for tendon healing: a systematic review. Br Med Bull. 2011;98:31-59.

Lochner K, Fritsche A, Jonitz A, Hansmann D, Mueller P, Mueller-Hilke B, Bader R. The potential role of human osteoblasts for periprosthetic osteolysis following exposure to wear particles. Int J Mol Med. 2011 Dec;28(6):1055-63.

Lohmann CH, Dean DD, Köster G, et al. Ceramic and PMMA particles differentially affect osteoblast phenotype. Biomaterials. 2002 Apr;23(8):1855-63.

Longo UG, Lamberti A, Maffulli N, Denaro V. Tissue engineered biological augmentation for tendon healing: a systematic review. Br Med Bull. 2011;98:31-59.

Lorenzo J, Horowitz M, Choi Y. Osteoimmunology: interactions of the bone and immune system. Endocr Rev. 2008 Jun;29(4):403-40.

Lories RJ, Derese I, Ceuppens JL, Luyten FP. Bone morphogenetic proteins 2 and 6, expressedinarthriticsynovium,areregulatedbyproinflammatorycytokinesanddifferentiallymodulatefibroblast-likesynoviocyteapoptosis.ArthritisRheum. 2003;48:2807-2818.

Lories RJ, Luyten FP. Bone morphogenetic protein signaling in joint homeostasis and disease. Cytokine Growth Factor Rev. 2005 Jun;16(3):287-98.

Lowery JW, de Caestecker MP. BMP signaling in vascular development and disease. Cytokine Growth Factor Rev. 2010 Aug;21(4):287-98.

Ma GK, Chiu R, Huang Z, Pearl J, Ma T, Smith RL, Goodman SB. Polymethyl-methacrylate particle exposure causes changes in p38 MAPK and TGF-beta signaling in differentiating MC3T3-E1 cells. J Biomed Mater Res A. 2010 Jul;94(1):234-40.

Madry H, Luyten FP, Facchini A. Biological aspects of early osteoarthritis. Knee Surg Sports Traumatol Arthrosc. 2012 Mar;20(3):407-22.

Maki RG. Small is beautiful: insulin-like growth factors and their role in growth, development, and cancer. J Clin Oncol. 2010 Nov 20;28(33):4985-95.

MaloneyWJ,JastyM,RosenbergA,HarrisWH.Bonelysisinwell-fixedcementedfemoral components. J Bone Joint Surg 1990;72B:966-70.

Maloney WJ, Peters P, Engh CA, Head WC, Chandler H. Severe pelvic osteolysis complicating cementless acetabular reconstruction. J Bone Joint Surg 1993;75A:1627-35.

Maloney WJ, Smith RL. Periprosthetic osteolysis in total hip arthroplasty: The role of particular wear debris. J Bone Joint Surg 1995;77A:1448-61.

91

Mandelin J, Li TF, Liljeström M, Kroon ME, Hanemaaijer R, Santavirta S, Konttinen YT. Imbalance of RANKL/RANK/OPG system in interface tissue in loosening of total hip replacement. J Bone Joint Surg Br. 2003 Nov;85(8):1196-201.

Mandelin J, Liljeström M, Li TF, Ainola M, Hukkanen M, Salo J, Santavirta S, Konttinen YT. Pseudosynovial fluid from loosened total hip prosthesisinduces osteoclast formation. J Biomed Mater Res B Appl Biomater. 2005a Jul;74(1):582-8.

Mandelin J, Li TF, Hukkanen M, Liljeström M, Salo J, Santavirta S, Konttinen YT.Interfacetissuefibroblastsfromloosetotalhipreplacementprosthesisproduce receptor activator of nuclear factor-kappaB ligand, osteoprotegerin, and cathepsin K. J Rheumatol. 2005b Apr;32(4):713-20.

ManlapazM,MaloneyWJ,SmithRL.Invitroactivationofhumanfibroblastsbyretrieved titanium alloy wear debris. J Orthop Res. 1996 May;14(3):465-72.

Markel DC, Zhang R, Shi T, Hawkins M, Ren W. Inhibitory effects of erythromycin onweardebris-inducedVEGF/Flt-1geneproductionandosteolysis.InflammRes. 2009 58:413-21.

Marshall A, Ries MD, Paprosky W; Implant Wear Symposium 2007 Clinical Work Group. How prevalent are implant wear and osteolysis, and how has the scope of osteolysis changed since 2000? J Am Acad Orthop Surg. 2008;16 Suppl 1:S1-6.

Martin TJ, Sims NA. Osteoclast-derived activity in the coupling of bone formation to resorption. Trends Mol Med. 2005 Feb;11(2):76-81.

Matthies AK, Skinner JA, Osmani H, Henckel J, Hart AJ. Pseudotumors are common in well-positioned low-wearing metal-on-metal hips. Clin Orthop Relat Res. 2012 Jul;470(7):1895-906.

Maurer T, Zimmermann G, Maurer S, Stegmaier S, Wagner C, Hänsch GM. Inhibition of osteoclast generation: a novel function of the bone morphogenetic protein 7/osteogenicprotein1.MediatorsInflamm.2012;2012:171209.

Mayr-Wohlfart U, Waltenberger J, Hausser H, et al. Vascular endothelial growth factor stimulates chemotactic migration of primary human osteoblasts. Bone. 2002 Mar;30(3):472-7.

McCarthy TL, Centrella M, Canalis E. Regulatory effects of insulin-like growth factors I and II on bone collagen synthesis in rat calvarial cultures. Endocrinology 1989;124: 301-9.

McEvoy A, Jeyam M, Ferrier G, Evans CE, Andrew JG. Synergistic effect of particles and cyclic pressure on cytokine production in human monocyte/macrophages: proposed role in periprosthetic osteolysis. Bone. 2002 Jan;30(1):171-7.

Middleton J, Arnott N, Walsh S, Beresford J. Osteoblasts and osteoclasts in adult human osteophyte tissue express the mRNAs for insulin-like growth factors I and II and the type 1 IGF receptor. Bone. 1995 Mar;16(3):287-93.

92

Min JK, Kim YM, Kim YM, et al. Vascular endothelial growth factor up-regulates expression of receptor activator of NF-kappa B (RANK) in endothelial cells. Concomitant increase of angiogenic responses to RANK ligand. J Biol Chem. 2003 Oct 10;278(41):39548-57.

Mishina Y, Starbuck MW, Gentile MA, et al. Bone morphogenetic protein type IA receptor signaling regulates postnatal osteoblast function and bone remodeling. J Biol Chem. 2004 Jun 25;279(26):27560-6.

Mirra JM, Marder RA, Amstutz HC. The pathology of failed total joint arthroplasty. Clin Orthop 1982;170:175-83.

Miyanishi K, Trindade MC, Ma T, Goodman SB, Schurman DJ, Smith RL. Periprosthetic osteolysis: induction of vascular endothelial growth factor from human monocyte/macrophages by orthopaedic biomaterial particles. J Bone Miner Res. 2003 Sep;18(9):1573-83.

Mochizuki H, Hakeda Y, Wakatsuki N, et al. Insulin-like growth factor-I supports formation and activation of osteoclasts. Endocrinology 1992;131:1075-80.

Mohan S, Bautista CM, Herring SJ, Linkhart TA, Baylink DJ. Development of valid methods to measure insulin-like growth factors-I and -II in bone cell-conditioned medium. Endocrinology. 1990 May;126(5):2534-42.

Mohan S, Baylink DJ. Bone growth factors. Clin Orthop Relat Res. 1991 Feb;(263):30-48.

Molloy T, Wang Y, Murrell G. The roles of growth factors in tendon and ligament healing. Sports Med. 2003;33(5):381-94.

Moester MJ, Papapoulos SE, Löwik CW, van Bezooijen RL. Sclerostin: current knowledge and future perspectives. Calcif Tissue Int. 2010 Aug;87(2):99-107.

Morrison DK, Kaplan DR, Rhee SG, Williams LT. Platelet-derived growth factor (PDGF)-dependent association of phospholipase C-gamma with the PDGF receptor signaling complex. Mol Cell Biol. 1990 May;10(5):2359-66.

Myllymaa S, Kaivosoja E, Myllymaa K, Sillat T, Korhonen H, Lappalainen R, Konttinen YT. Adhesion, spreading and osteogenic differentiation of mesenchymal stem cells cultured on micropatterned amorphous diamond, titanium, tantalum and chromium coatings on silicon. J Mater Sci Mater Med. 2010 Jan;21(1):329-41.

Mäkelä KT, Eskelinen A, Pulkkinen P, Virolainen P, Paavolainen P, Remes V. Cementedversuscementless totalhipreplacements inpatientsfifty-fiveyears of age or older with rheumatoid arthritis. J Bone Joint Surg Am. 2011 Jan 19;93(2):178-86. (2011a)

Mäkelä KT, Eskelinen A, Pulkkinen P, Paavolainen P, Remes V. Results of 3,668 primary total hip replacements for primary osteoarthritis in patients under the age of 55 years. Acta Orthop. 2011 Oct;82(5):521-9. (2011b)

NabaeM,InoueK,UshiyamaT,HukudaS.Geneexpressionsofantiinflammatory

93

mediators in THR retrieved interfacial membranes. Acta Orthop Scand. 1999 Apr;70(2):149-54.

Nakagawa N, Kinosaki M, Yamaguchi K, et al. RANK is the essential signaling receptor for osteoclast differentiation factor in osteoclastogenesis. Biochem Biophys Res Commun. 1998 Dec 18;253(2):395-400.

Nakagawa M, Kaneda T, Arakawa T, et al. Vascular endothelial growth factor (VEGF) directly enhances osteoclastic bone resorption and survival of mature osteoclasts. FEBS Lett. 2000 May 12;473(2):161-4.

Nakanishi Y, Kihara K, Mizuno K, Masamune Y, Yoshitake Y, Nishikawa K. Direct effectofbasicfibroblastgrowthfactorongenetranscriptioninacell-freesystem. Proc Natl Acad Sci U S A. 1992;89(12):5216-20.

Nash TJ, Howlett CR, Martin C, Steele J, Johnson KA, Hicklin DJ. Effect of platelet-derived growth factor on tibial osteotomies in rabbits. Bone. 1994 Mar-Apr;15(2):203-8.

Nathan C, Sporn M. Cytokines in context. J Cell Biol 1991;113:981-6.National Collaborating Centre for Chronic Conditions. Osteoarthritis: national

clinical guideline for care and management in adults. London: Royal College of Physicians, 2008. http://www.nice.org.uk/nicemedia/pdf/CG059FullGuideline.pdf

Neufeld G, Cohen T, Gengrinovitch S, Poltorak Z. Vascular endothelial growth factor (VEGF) and its receptors. FASEB J. 1999 Jan;13(1):9-22.

Nicolas V, Prewett A, Bettica P, Mohan S, Finkelman RD, Baylink DJ, Farley JR. Age-related decreases in insulin-like growth factor-I and transforming growth factor-beta in femoral cortical bone from both men and women: implications for bone loss with aging. J Clin Endocrinol Metab. 1994 May;78(5):1011-6.

Niida S, Kaku M, Amano H, et al. Vascular endothelial growth factor can substitute for macrophage colony-stimulating factor in the support of osteoclastic bone resorption. J Exp Med. 1999 Jul 19;190(2):293-8.

Niissalo S, Li TF, Santavirta S, Takagi M, Hietanen J, Konttinen YT. Dense innervation in pseudocapsular tissue compared to aneural interface tissue in loose totally replaced hips. J Rheumatol. 2002 Apr;29(4):796-803.

Nordsletten L, Buo L, Takagi M, Konttinen YT, Yamakawa M, Santavirta S, Aasen AO. The plasminogen activation system is upregulated in loosening of total hip prostheses. Acta Orthop Scand. 1996 Apr;67(2):143-8.

Ogino Y, Ayukawa Y, Kukita T, Koyano K. The contribution of platelet-derived growth factor, transforming growth factor-beta1, and insulin-like growth factor-I in platelet-rich plasma to the proliferation of osteoblast-like cells. Oral Surg Oral Med Oral Pathol Oral Radiol Endod. 2006 Jun;101(6):724-9.

94

Oh JS, Kucab JE, Bushel PR, et al. Insulin-like growth factor-1 inscribes a gene expressionprofileforangiogenicfactorsandcancerprogressioninbreastepithelial cells. Neoplasia. 2002 May-Jun;4(3):204-17.

Ohlin A, Johnell O, Lerner UH. The pathogenesis of loosening of total hip arthroplasties: the production of factors by periprosthetic tissues that stimulate in vitro bone resorption. Clin Orthop 1990;253:287-96.

Ohlin A, Lerner UH. Bone-resorbing activity of different periprosthetic tissues in aseptic loosening of total hip arthroplasty. Bone Miner 1993;20(1):67-78.

Ohlsson C, Bengtsson BA, Isaksson OG, Andreassen TT, Slootweg MC. Growth hormone and bone. Endocr Rev. 1998;19(1):55-79.

Pakvis D, van Hellemondt G, de Visser E, Jacobs W, Spruit M. Is there evidence forasuperiormethodofsocketfixationinhiparthroplasty?Asystematicreview. Int Orthop. 2011 Aug;35(8):1109-18.

PapT,ClausA,OhtsuS,etal.Osteoclast-independentboneresorptionbyfibroblast-like cells. Arthritis Res Ther. 2003;5(3):R163-73.

Parikh AA, Ellis LM. The vascular endothelial growth factor family and its receptors. Hematol Oncol Clin North Am. 2004 Oct;18(5):951-71.

ParkJB.Effectsoffibroblastgrowth factor2onosteoblasticproliferationanddifferentiation by regulating bone morphogenetic protein receptor expression. J Craniofac Surg. 2011 Sep;22(5):1880-2.

Pastides PS, Khan W. Rotator cuff injuries: the evolving role of tissue engineering. British Journal of Medicine & Medical Research 2012;2(1):1-14.

Pauly S, Klatte F, Strobel C, Schmidmaier G, Greiner S, Scheibel M, Wildemann B. BMP-2 and BMP-7 affect human rotator cuff tendon cells in vitro. J Shoulder Elbow Surg. 2012 Apr;21(4):464-73.

Pederson L, Ruan M, Westendorf JJ, Khosla S, Oursler MJ. Regulation of bone formation by osteoclasts involves Wnt/BMP signaling and the chemokine sphingosine-1-phosphate. Proc Natl Acad Sci U S A. 2008 Dec 30;105(52):20764-9.

Pettit AR, Chang MK, Hume DA, Raggatt LJ. Osteal macrophages: a new twist on coupling during bone dynamics. Bone. 2008 Dec;43(6):976-82.

Perry MJ, Mortuza FY, Ponsford FM, Elson CJ, Atkins RM, Learmonth ID. Properties of tissue from around cemented joint implants with erosive and/or linear osteolysis. J Arthroplasty. 1997 Sep;12(6):670-6.

Pilkington MF, Sims SM, Dixon SJ. Transforming growth factor-beta induces osteoclastrufflingandchemotaxis:potentialroleinosteoclastrecruitment.J Bone Miner Res. 2001 Jul;16(7):1237-47.

Prabhu A, Shelburne CE, Gibbons DF. Cellular proliferation and cytokine responses of murine macrophage cell line J774A.1 to polymethylmethacrylate and cobalt-chrome alloy particles. J Biomed Mater Res. 1998 Dec 15;42(4):655-63.

95

Purdue PE, Koulouvaris P, Potter HG, Nestor BJ, Sculco TP. The cellular and molecular biology of periprosthetic osteolysis. Clin Orthop Relat Res. 2007;454:251-61.

Quinn JM, Elliott J, Gillespie MT, Martin TJ. A combination of osteoclast differentiationfactorandmacrophage-colonystimulatingfactorissufficientfor both human and mouse osteoclast formation in vitro. Endocrinology. 1998 Oct;139(10):4424-7.

Quinn JM, Itoh K, Udagawa N, et al. Transforming growth factor beta affects osteoclast differentiation via direct and indirect actions. J Bone Miner Res. 2001 Oct;16(10):1787-94.

Rader ChP, Baumann B, Rolf O, et al. Detection of differentially expressed genes inparticlediseaseusingarray-filteranalysis.BiomedTech (Berl).2002May;47(5):111-6.

Ramoshebi LN, Ripamonti U. Osteogenic protein-1, a bone morphogenetic protein, induces angiogenesis in the chick chorioallantoic membrane and synergizes withbasicfibroblastgrowthfactorandtransforminggrowthfactor-beta1.Anat Rec. 2000 May 1;259(1):97-107.

RenR,CharlesPC,ZhangC,WuY,WangH,PattersonC.Geneexpressionprofilesidentify a role for cyclooxygenase 2-dependent prostanoid generation in BMP6-induced angiogenic responses. Blood. 2007 Apr 1;109(7):2847-53.

Re RN, Cook JL. Senescence, apoptosis, and stem cell biology: the rationale for an expanded view of intracrine action. Am J Physiol Heart Circ Physiol. 2009;297(3):H893-901.

Ren WP, Markel DC, Zhang R, Peng X, Wu B, Monica H, Wooley PH. Association between UHMWPE particle-induced inflammatory osteoclastogenesis and expression of RANKL, VEGF, and Flt-1 in vivo. Biomaterials. 2006 Oct;27(30):5161-9.

Ren W, Zhang R, Markel DC, Wu B, Peng X, Hawkins M, Wooley PH. Blockade of vascular endothelial growth factor activity suppresses wear debris-induced inflammatoryosteolysis.JRheumatol.2007Jan;34(1):27-35.

Ren W, Zhang R, Wu B, Wooley PH, Hawkins M, Markel DC. Effects of SU5416 and a vascular endothelial growth factor neutralizing antibody on wear debris-inducedinflammatoryosteolysisinamousemodel.JInflammRes.2011;4:29-38.

Rader ChP, Baumann B, Rolf O, et al. Detection of differentially expressed genes inparticlediseaseusingarray-filteranalysis.BiomedTech (Berl).2002May;47(5):111-6.

Raggatt LJ, Partridge NC. Cellular and molecular mechanisms of bone remodeling. J Biol Chem. 2010 Aug 13;285(33):25103-8.

96

Revell PA. The combined role of wear particles, macrophages and lymphocytes in the loosening of total joint prostheses. J R Soc Interface. 2008 Nov 6;5(28):1263-78.

Rodeo SA. Biologic augmentation of rotator cuff tendon repair. J Shoulder Elbow Surg. 2007 Sep-Oct;16(5 Suppl):S191-7.

Rolf O, Baumann B, Sterner T, et al. Characterization of mode II-wear particles and cytokine response in a human macrophage-like cell culture. Biomed Tech (Berl). 2005 Jan-Feb;50(1-2):25-9.

Rosen V, Wozney JM. Bone morphogenetic proteins. In: Bilezikian JP, Raisz LG, Rodan GA, eds. Vol 2. 2nd ed. San Diego: Academic Press; 2002:919-928.

Sabokbar A, Fujikawa Y, Neale S, Murray DW, Athanasou NA. Human arthroplasty derived macrophages differentiate into osteoclastic bone resorbing cells. Ann Rheum Dis. 1997 Jul;56(7):414-20.

SaitoI,KoshinoT,NakashimaK,UesugiM,SaitoT.Increasedcellularinfiltratein inflammatorysynoviaofosteoarthriticknees.OsteoarthritisCartilage.2002 Feb;10(2):156-62.

Sakai H, Jingushi S, Shuto T, et al. Fibroblasts from the inner granulation tissue of the pseudocapsule in hips at revision arthroplasty induce osteoclast differentiation, as do stromal cells. Ann Rheum Dis. 2002 Feb;61(2):103-9.

Salo J, Lehenkari P, Mulari M, Metsikkö K, Väänänen HK. Removal of osteoclast bone resorption products by transcytosis. Science. 1997 Apr 11;276(5310):270-3.

Sanchez-FernandezMA,GalloisA,RiedlT,JurdicP,HoflackB.Osteoclastscontrolosteoblast chemotaxis via PDGF-BB/PDGF receptor beta signaling. PLoS One. 2008;3(10):e3537.

SanjabiS,ZenewiczLA,KamanakaM,FlavellRA.Anti-inflammatoryandpro-inflammatory roles of TGF-beta, IL-10, and IL-22 in immunity and autoimmunity. Curr Opin Pharmacol. 2009 Aug;9(4):447-53.

Santavirta S, Konttinen YT, Bergroth V, Eskola A, Tallroth K, Lindholm TS. Aggressive granulomatous lesions associated with hip arthroplasty. Immunopathological studies. J Bone Joint Surg Am. 1990a Feb;72(2):252-8.

Santavirta S, Hoikka V, Eskola A, Konttinen YT, Paavilainen T, Tallroth K. Aggressive granulomatous lesions in cementless total hip arthroplasty. J Bone Joint Surg Br. 1990b Nov;72(6):980-4.

Santavirta S, Konttinen YT, Hoikka V, Eskola A. Immunopathological response to loose cementless acetabular components. J Bone Joint Surg 1991;73B:38-42.

Santavirta S, Sorsa T, Konttinen YT, Saari H, Eskola A, Eisen A. Role of mesenchymal collagenase in the loosening of total hip prosthesis. Clin Orthop 1993;290:206-15.

97

Santavirta S, Ceponis A, Solovieva SA, Hurri H, Jin J, Takagi M, Suda A, Konttinen YT. Periprosthetic microvasculature in loosening of total hip replacement. Arch Orthop Trauma Surg. 1996;115(5):286-9.

Santavirta S, Xu JW, Hietanen J, Ceponis A, Sorsa T, Kontio R, Konttinen YT. Activation of periprosthetic connective tissue in aseptic loosening of total hip replacements. Clin Orthop Relat Res. 1998 Jul;(352):16-24.

Sara VR, Hall K. Insulin-like growth factors and their binding proteins. Physiol Rev. 1990 Jul;70(3):591-614.

Sato J, Segami N, Yoshitake Y, Nishikawa K. Correlations of the expression of fibroblast growth factor-2, vascular endothelial growth factor, andtheir receptors with angiogenesis in synovial tissues from patients with internal derangement of the temporomandibular joint. J Dent Res. 2003 Apr;82(4):272-7.

Sawano A, Iwai S, Sakurai Y, Ito M, Shitara K, Nakahata T, Shibuya M. Flt-1, vascular endothelial growth factor receptor 1, is a novel cell surface marker for the lineage of monocyte-macrophages in humans. Blood. 2001 Feb 1;97(3):785-91.

Schmalzried TP, Kwong LM, Jasty M. The mechanism of loosening of cemented acetabular components in total hip arthroplasty: analysis of specimens retrieved at autopsy. Clin Orthop 1992a;274:60-78.

Schmalzried TP, Jasty M, Harris WH. Periprosthetic bone loss in total hip arthroplasty. Polyethylene wear debris and the concept of the effective joint space. J Bone Joint Surg Am. 1992b Jul;74(6):849-63.

Scott WW, Riley LH, Dorfman HD. Focal lytic lesions associated with femoral stem loosening in total hip prosthesis. Am J Radiol 1985;144:977-82.Shanbhag AS, Jacobs JJ, Black J, Galante JO, Glant TT. Cellular mediators secreted by interfacial membranes obtained at revision total hip arthroplasty. J Arthroplasty 1995;10:498-0506.

Seghezzi G, Patel S, Ren CJ, et al. Fibroblast growth factor-2 (FGF-2) induces vascular endothelial growth factor (VEGF) expression in the endothelial cells of forming capillaries: an autocrine mechanism contributing to angiogenesis. J Cell Biol. 1998 Jun 29;141(7):1659-73.

Shanbhag AS, Jacobs JJ, Black J, Galante JO, Glant TT. Cellular mediators secreted by interfacial membranes obtained at revision total hip arthroplasty. J Arthroplasty. 1995 Aug;10(4):498-506.

Shanbhag AS, Kaufman AM, Hayata K, Rubash HE. Assessing osteolysis with use of high-throughput protein chips. J Bone Joint Surg Am. 2007 May;89(5):1081-9.

Shao ES, Lin L, Yao Y, Boström KI. expression of vascular endothelial growth factor is coordinately regulated by the activin-like kinase receptors 1 and 5 in endothelial cells. Blood. 2009 Sep 3;114(10):2197-206.

98

Shimoaka T, Ogasawara T, Yonamine A, et al. Regulation of osteoblast, chondrocyte, andosteoclastfunctionsbyfibroblastgrowthfactor(FGF)-18incomparisonwith FGF-2 and FGF-10. J Biol Chem. 2002 Mar 1;277(9):7493-500.

SHPR (Svenska höftprotesregistret). Swedish hip arthroplasty register 2011. 2012 :1-136. http://www.shpr.se/Files/Årsrapport%202011%20(eng)%20webb.pdf

Simonet WS, Lacey DL, Dunstan CR, et al. Osteoprotegerin: a novel secreted protein involved in the regulation of bone density. Cell. 1997 Apr 18;89(2):309-19.

Spanogle JP, Miyanishi K, Ma T, Epstein NJ, Smith RL, Goodman SB. Comparison of VEGF-producing cells in periprosthetic osteolysis. Biomaterials. 2006 Jul;27(21):3882-7.

STAKES.Lonkka-japolviproteesitSuomessa2011.2013Sept:1-22.http://urn.fi/URN:NBN:fi-fe201303222594

Steffens GC, Yao C, Prével P, Markowicz M, Schenck P, Noah EM, Pallua N. Modulation of angiogenic potential of collagen matrices by covalent incorporation of heparin and loading with vascular endothelial growth factor. Tissue Eng. 2004 Sep-Oct;10(9-10):1502-9.

Street J, Bao M, deGuzman L, et al. Vascular endothelial growth factor stimulates bone repair by promoting angiogenesis and bone turnover. Proc Natl Acad Sci U S A. 2002 Jul 23;99(15):9656-61.

Street J, Lenehan B. Vascular endothelial growth factor regulates osteoblast survival - evidence for an autocrine feedback mechanism. J Orthop Surg Res. 2009 Jun 16;4:19.

Stulberg SD,WixsonRL, AdamsAD,HendrixRW,Bernfield JB.Monitoringpelvic osteolysis following total hip replacement surgery: an algorithm for surveillance. J Bone Joint Surg 2002;84A(Suppl):116–122.

SuEJ,CioffiCL,StefanssonS,MitterederN,GarayM,HreniukD,LiauG.Genetherapy vector-mediated expression of insulin-like growth factors protects cardiomyocytes from apoptosis and enhances neovascularization. Am J Physiol Heart Circ Physiol. 2003 Apr;284(4):H1429-40.

SunJS,LinFH,HungTY,TsuangYH,ChangWH,LiuHC.The influenceofhydroxyapatite particles on osteoclast cell activities. J Biomed Mater Res. 1999 Jun 15;45(4):311-21.

Takagi M, Konttinen YT, Santavirta S, Sorsa T, Eisen AZ, Nordsletten L, Suda A. Extracellular matrix metalloproteinases around loose total hip prostheses. Acta Orthop Scand. 1994 Jun;65(3):281-6.

Takagi M, Santavirta S, Ida H, Ishii M, Takei I, Niissalo S, Ogino T, Konttinen YT. High-turnover periprosthetic bone remodeling and immature bone formation around loose cemented total hip joints. J Bone Miner Res. 2001 Jan;16(1):79-88.

99

Takai H, Kanematsu M, Yano K, et al. Transforming growth factor-beta stimulates the production of osteoprotegerin/osteoclastogenesis inhibitory factor by bone marrow stromal cells. J Biol Chem. 1998 Oct 16;273(42):27091-6.

Takei I, Takagi M, Ida H, Ogino T, Santavirta S, Konttinen YT. High macrophage-colonystimulatingfactorlevelsinsynovialfluidoflooseartificialhipjoints.J Rheumatol. 2000 Apr;27(4):894-9.

Thorey F, Weinert K, Weizbauer A, et al. Coating of titanium implants with copolymer supports bone regeneration: a comparative in vivo study in rabbits. J Appl Biomater Biomech. 2011 Jan-Apr;9(1):26-33.

TsaiJA,AnderssonMK,IvarssonM,GranbergB,StarkA.Effectsofsynovialfluidfrom aseptic prosthesis loosening on collagen production in osteoblasts. Int Orthop. 2009 Jun;33(3):873-7.

Tsiridis E, Upadhyay N, Giannoudis P. Molecular aspects of fracture healing: which are the important molecules? Injury. 2007 Mar;38 Suppl 1:S11-25.

Tuan RS, Lee FY, T Konttinen Y, Wilkinson JM, Smith RL; Implant Wear Symposium2007 Biologic Work Group. What are the local and systemic biologic reactions

and mediators to wear debris, and what host factors determine or modulate the biologic response to wear particles? J Am Acad Orthop Surg. 2008;16 Suppl 1:S42-8.

Tunyogi-Csapo M, Koreny T, Vermes C, Galante JO, Jacobs JJ, Glant TT. Role of fibroblasts and fibroblast-derived growth factors in periprosthetic angiogenesis. J Orthop Res. 2007 Oct;25(10):1378-88.

Ulrich-Vinther M. Gene therapy methods in bone and joint disorders. Evaluation of the adeno-associated virus vector in experimental models of articular cartilage disorders, periprosthetic osteolysis and bone healing. Acta Orthop Suppl. 2007 Apr;78(325):1-64.

Ucuzian AA, Gassman AA, East AT, Greisler HP. Molecular mediators of angiogenesis. J Burn Care Res. 2010 Jan-Feb;31(1):158-75.

Urist MR. Bone: formation by autoinduction. Science. 1965 Nov 12;150(3698):893-9.van Bezooijen RL, Roelen BA, Visser A, et al. Sclerostin is an osteocyte-expressed

negative regulator of bone formation, but not a classical BMP antagonist. J Exp Med. 2004 Mar 15;199(6):805-14.

van der Horst G, van Bezooijen RL, Deckers MM, et al. Differentiation of murine preosteoblastic KS483 cells depends on autocrine bone morphogenetic protein signaling during all phases of osteoblast formation. Bone. 2002;31(6):661-9.

Venesmaa P, Vanninen E, Miettinen H, Kröger H. Periprosthetic bone turnover after primary total hip arthroplasty measured by single-photon emission computed tomography. Scand J Surg. 2012;101(4):241-8.

100

Välimäki VV, Yrjans JJ, Vuorio E, Aro HT. Combined effect of BMP-2 gene transfer and bioactive glass microspheres on enhancement of new bone formation. J Biomed Mater Res A. 2005 Dec 1;75(3):501-9.

Wan Z, Dorr LD. Natural history of femoral focal osteolysis with proximal ingrowth smooth stem implant. J Arthroplasty. 1996 Sep;11(6):718-25.

Wang CT, Lin YT, Chiang BL, Lee SS, Hou SM. Over-expression of receptor activator ofnuclear factor-kappaB ligand (RANKL), inflammatory cytokines, andchemokines in periprosthetic osteolysis of loosened total hip arthroplasty. Biomaterials. 2010 Jan;31(1):77-82.

Wang H, Jia TH, Zacharias N, Gong W, Du HX, Wooley PH, Yang SY. Combination genetherapytargetingoninterleukin-1βandRANKLforweardebris-inducedaseptic loosening. Gene Ther. 2013 Feb;20(2):128-35.

Wang JS. Basic fibroblast growth factor for stimulation of bone formation inosteoinductive or conductive implants. Acta Orthop Scand Suppl. 1996 Apr;269:1-33.

Wang JY, Wicklund BH, Gustilo RB, Tsukayama DT. Titanium, chromium and cobalt ions modulate the release of bone-associated cytokines by human monocytes/macrophages in vitro. Biomaterials. 1996a Dec;17(23):2233-40.

Wang JY, Tsukayama DT, Wicklund BH, Gustilo RB. The increase in bone-resorbing cytokinesinthejointfluidsandserainpatientswithtotaljointarthroplasty.Trans Orthop Res Soc 1996b;275.

Wang ML, Sharkey PF, Tuan RS. Particle bioreactivity and wear-mediated osteolysis. J Arthroplasty. 2004 Dec;19(8):1028-38.

Wang W, Ferguson DJ, Quinn JM, Simpson AH, Athanasou NA. Biomaterial particle phagocytosis by bone-resorbing osteoclasts. J Bone Joint Surg Br. 1997 Sep;79(5):849-56.

Wang W, Ouyang Y, Poh CK. Orthopaedic implant technology: biomaterials from past to future. Ann Acad Med Singapore. 2011 May;40(5):237-8.

WangQ,RozelleAL,LepusCM,etal.Identificationofacentralroleforcomplementin osteoarthritis. Nat Med 2011;17:1674–9.

WarisV,XuJW,NordslettenL,SorsaT,SantavirtaS,KonttinenYT.Basicfibroblastgrowth factor (bFGF) in the synovial-like membrane around loose total hip prostheses. Scand J Rheumatol. 1996;25(4):257-62.

Waris V, Zhao DS, Leminen H, Santavirta S, Takagi M, Nordsletten L, Konttinen YT. Insulin-like growth factors I and II in the aseptic loosening of total hip implants. Scand J Rheumatol. 2004;33(6):428-31.

Waris V, Waris E, Sillat T, Konttinen YT. BMPs in periprosthetic tissues around aseptically loosened total hip implants. Acta Orthop. 2010 Aug;81(4):420-6.

101

Willert HG, Ludwig J, Semlitsch M. Reaction of bone to methacrylate after hip arthroplasty: a long-term gross, light microscopic, and scanning electron microscopic study. J Bone Joint Surg 1974;56A:1368-82.

Willert HG, Semlitsch M. Tissue reactions to plastic and metallic wear products of joint endoprostheses. In: Geschwind N, Debrunner HU, eds. Total hip prosthesis. Baltimore: Williams and Wilkins, 1976;205-39.

Willert HG, Semlitsch M. Reactions of the articular capsule to wear products of artificialjointprosthesis.JBiomedMaterRes1977;11:157-64.

Willert HG, Bertram H, Buchhom GH. Osteolysis in alloarthroplasty of the hip. The role of ultrahigh molecular weight polyethylene wear particles. Clin Orthop 1990;258:95-107.

Xu JW, Konttinen YT, Waris V, Pätiälä H, Sorsa T, Santavirta S. Macrophage-colony stimulating factor (M-CSF) is increased in the synovial-like membrane of the periprosthetic tissues in the aseptic loosening of total hip replacement (THR). Clin Rheumatol. 1997 May;16(3):243-8.

Xu JW, Konttinen YT, Li TF, Waris V, Lassus J, Matucci-Cerinic M, Sorsa T, Santavirta TS. Production of platelet-derived growth factor in aseptic loosening of total hip replacement. Rheumatol Int. 1998;17(6):215-21.

Xu JW, Li TF, Partsch G, Ceponis A, Santavirta S, Konttinen YT. Interleukin-11 (IL-11) in aseptic loosening of total hip replacement (THR). Scand J Rheumatol. 1998;27(5):363-7.

Youinou P, Jamin C. The weight of interleukin-6 in B cell-related autoimmune disorders. J Autoimmun. 2009 May-Jun;32(3-4):206-10.

Yu X, Huang Y, Collin-Osdoby P, Osdoby P. Stromal cell-derived factor-1 (SDF-1) recruits osteoclast precursors by inducing chemotaxis, matrix metalloproteinase-9 (MMP-9) activity, and collagen transmigration. J Bone Miner Res. 2003 Aug;18(8):1404-18.

Zelzer E, Olsen BR. Multiple roles of vascular endothelial growth factor (VEGF) in skeletal development, growth, and repair. Curr Top Dev Biol. 2005;65:169-87.

Zhang W, Peng X, Cheng T, Zhang X. Vascular endothelial growth factor gene silencingsuppressesweardebris-inducedinflammation.IntOrthop.2011Dec;35(12):1883-8.

Zhang Z, Chen J, Jin D. Platelet-derived growth factor (PDGF)-BB stimulates osteoclastic bone resorption directly: the role of receptor beta. Biochem Biophys Res Commun. 1998 Oct 9;251(1):190-4.

Zhao C, Irie N, Takada Y, et al. Bidirectional ephrinB2-EphB4 signaling controls bone homeostasis. Cell Metab. 2006 Aug;4(2):111-21.

Zreiqat H, Kumar RK, Markovic B, Zicat B, Howlett CR. Macrophages at the skeletal tissue-device interface of loosened prosthetic devices express bone-related genes and their products. J Biomed Mater Res A. 2003 Apr 1;65(1):109-17.

102

Yang PJ, Temenoff JS. Engineering orthopedic tissue interfaces. Tissue Eng Part B Rev. 2009 Jun;15(2):127-41.

YaoJ,GlantTT,LarkMW,etal.Thepotentialroleoffibroblastsinperiprostheticosteolysis:fibroblastresponsetotitaniumparticles.JBoneMinerRes.1995Sep;10(9):1417-27.

Yasuda H, Shima N, Nakagawa N, et al. Osteoclast differentiation factor is a ligand for osteoprotegerin/osteoclastogenesis-inhibitory factor and is identical to TRANCE/RANKL. Proc Natl Acad Sci U S A. 1998 Mar 31;95(7):3597-602.

Yoshida H, Hayashi S, Kunisada T, et al. The murine mutation osteopetrosis is in the coding region of the macrophage colony stimulating factor gene. Nature. 1990 May 31;345(6274):442-4.

Yun YR, Won JE, Jeon E, Lee S, Kang W, Jo H, Jang JH, Shin US, Kim HW. Fibroblast growth factors: biology, function, and application for tissue regeneration. J Tissue Eng. 2010 Nov 7;2010:218142.

103

ORIGINAL PUbLICATIONS