Research Collection

Doctoral Thesis

Differentially load-regulated gene expression in mouse trabecular osteocytes

Author(s): Wasserman, Elad

Publication Date: 2010

Permanent Link: https://doi.org/10.3929/ethz-a-006128911

Rights / License: In Copyright - Non-Commercial Use Permitted

This page was generated automatically upon download from the ETH Zurich Research Collection. For more information please consult the Terms of use.

ETH Library DISS. ETH No. 18938

Differentially Load-Regulated Gene Expression in Mouse Trabecular Osteocytes

A dissertation submitted to ETH Zurich

for the degree of Doctor of Sciences

presented by

ELAD WASSERMAN

M.Sc. Cell Biology and Histology, Tel-Aviv University

born 27th September, 1972

citizen of Israel

accepted on the recommendation of

Examiner: Prof. Dr. Ralph Müller Co-Examiner: Prof. Dr. Itai Bab

2010

Table of contents:

Acknowledgements………………………………………………………………………………….v

Summary……………………………………………………………………………………………vi

Zusammenfassung………………………………………………………………………………….ix

1. Introduction……………………………………………………………………………………...13

1.1. Hypotheses and specific aims……………………………………………………………….15

1.1.1. Developing method of isolating mRNA from trabecular osteocytes…………………..16

1.1.2. Identify load-induced differentially regulated genes…………………………………..16

1.2. Outline of the thesis…………………………………………………………………………17

2. Background……………………………………………………………………………………...21

2.1. Bone anatomy……………………………………………………………………………….22

2.2. Osteocytes as mechanosensors/endocrine and paracrine organ …………………………….26

2.3. Bone remodeling…………………………………………………………………………….33

2.4. Mechanical load-induced bone adaptation .…………………………………..…………….40

2.5. Mouse genetics……………………………………………………………………………...45

3. Developing a method for isolation of osteocyte RNA………………………………………....67

3.1. Separation of trabecular bone from caudal vertebra………………………………………...68

3.2. Enzymatic digestion of non-osteocytic cells………………………………………………..70

3.3. RNA extraction from denuded trabeculae…………………………………………………..71

3.4. Comparative marker gene mRNA expression in enzymatically isolated cell fractions

and extracted RNA from denuded trabecular bone…………………………………………75

4. Load-induced differential regulation mRNA of trabecular osteocytes……………………...89

4.1. Single loading………………………………………………………………………………89

4.2. Repetitive loading…………………………………………………………………………..99 4.3. Functional genomics for identification of load-regulated pathways……………………...107

4.4. Confirmation of individual load-regulated genes in single loading………………………116

5. Synthesis………………………………………………………………………………………..127

Appendix…………………………………………………………………………………………..133

A1-A4. List of up- and down-regulated genes………………………………………………...134

A5-A6. List of load-regulated signalling pathways………………………………………...... 184

Curriculum Vitae ………………………………………………………………………………...190

Acknowledgements

The work presented in this thesis is the direct result of a great team effort and I am gratefully indebted to every member of that team.

First of all I would like to express my deep gratitude to both my supervisors, Professor Dr. Ralph Müller and Professor Dr. Itai Bab. Without their passion, guidance and leadership the successful realization of this project would not have been possible. Furthermore, I would like to thank them for their time, and the open door policy for which I am eternally grateful. I would also like to thank PD Dr. Franz Weber (University of Zürich) and Dr. Haike Hall-Bozic for their support and for providing me facilities in their laboratories.

Special thanks go to Dr. Duncan Webster for all of his help in doing the many experiments. His company and expertise were invaluable and significantly contributed to the outcome of this thesis. I also would like to thank Dr. Gisela Kuhn and Floor Lambers for their help during loading studies.

During my thesis, I enjoyed a fruitful close collaboration with The Functional Genomics Center Zurich (FGCZ) located at the University of Zurich. Many members of this institute made it possible to perform differential expression microarrays and bioinformatics statistical analyses.

Particular thanks go to the past and present members of the Institute of Biomechanics. Their expertise and companionship created a pleasant environment in which to work. I sincerely hope to stay in contact and share some good times with you in the future.

I am deeply grateful to my parents. Their unconditional love and support have enabled me to achieve all my goals. These few words cannot even begin to describe my gratitude and appreciation for all they have done for me.

Finally, the financial support of the Swiss National Science Foundation (SNF) and the Swiss Federal Institute of Technology (ETHZ) is gratefully acknowledged.

- v- Summary In light of the many bone diseases and injuries for which proper treatment is yet to be developed, there is a significant need to further address new approaches to stimulate bone healing. Osteoporosis is a disease characterized by an excessive decrease in bone mass which can lead to an increased susceptibility to fractures, skeletal deformation and, in more severe cases, death owing to morbidity. The disease has been attributed to both genetic and age-related factors. There are various medications available, which have been shown to delay bone loss, however no cure is yet achievable. To treat the disease, medical research is attempting to target genes which define osteoporosis, using the mouse as a model. Owing to the recent deciphering of the mouse genome and the high homology that exists between human and mouse genomes, inbred strains of mice represent ideal models for genetic studies. Using the mouse to identify genes implicated in the bone remodeling process could lead to advances in understanding that enable the precise regulation of genes and proteins responsible for particular bone phenotypes, i.e. bone mineral density or bone strength. One interesting phenotype under investigation is the response of bone to mechanical loading or its “mechano-sensitivity”. Mechanical loading is perhaps the most important single physiological/environmental factor regulating bone mass and shape. Age-related bone loss and consequent osteoporosis have been attributed, at least in part, to a reduction in muscle mass/function and the resultant decrease in mechanical usage of the skeleton. On the other hand, mechanical overloading has been shown to enhance bone formation and cause a net gain in cancellous bone mass, the major structural component of skeletal load-bearing sites. However, very little is known about the mechanisms involved in the load-induced anabolic effects in trabecular bone, mainly due to the lack of in vivo models to study load-induced molecular events. Building on the studies investigating the effect of mechanical loading on trabecular bone adaptation in the C56BL/6 mouse tail model that was recently developed in our group, the next long-term objective of this thesis was to elucidate the molecular mechanisms involved in the osteogenic anabolic effect of mechanical loading and to find genes and gene pathways that are regulated by mechanical loading. Mice have a well-characterized genome accessible to manipulations by transgenic and knockout technologies. An understanding of the molecular pathways governing load-stimulated bone formation could provide opportunities to mimic or augment bone mechano-sensitivity using pharmacological and molecular agents thereby leading to the development of novel strategies in the management of osteoporosis and other skeletal deficits.

- vi- To investigate the genetic regulation of mechanical loading, an ex-vivo method was established for the isolation of representative samples of mouse vertebral intact ribonucleic acid (RNA) derived selectively from trabecular osteoblast/lining cells and osteocytes, by using sequential collagenase digestions and pulverization. High quality total RNA preparations were isolated immediately following cell separation using conventional reagents and protocols. The quantity of total RNA isolated from trabecular osteocytes of a single caudal vertebra was sufficient for further differential gene expression analysis. To investigate a single and repetitive mechanical load-induced differential gene expression, the fifth caudal vertebra (C5) of C57BL/6 (B6) female mice was mechanically stimulated by respective single or repetitive load doses, each dose consisting of 3’000 cycles at a frequency of 10 Hz with an amplitude of 0N and 8N via two pins inserted into the adjacent vertebrae (C4 and C6). Mice were sacrificed six hours after the last mechanical loading and high quality total RNA preparations were analyzed for gene expression arrays, using Affymethrix Mouse Genome chips. Differential gene expression analysis of a single mechanical load revealed a total of 331 significantly regulated genes (P < 0.05), including 281 up-regulated probes and 50 down-regulated probes. Also, functional genomics analysis of acute loading, using GeneGo MetaCore software, indicated 65 load-regulated molecular pathways in which significantly regulated probes were present. In particular, up- regulation of insulin growth factor 1 (IGF-1, 2.2 fold) and wingless-type MMTV integration site family, member 5a (Wnt5a, 3.4 fold) genes have been shown which are thought to be activators of osteoblasts differentiation. In contrast, down-regulation of WNT inhibitor factor 1 gene (WIF-1, 1.8 fold) was shown, an inhibitor of WNT/beta-cathenin pathway for activation of osteoblast differentiation. Differential gene expression analysis of repetitive mechanical loading (three times per week of single load over four weeks) has shown a total of 1342 significantly regulated probes, involving 781 up-regulated and 561 down-regulated genes. In addition, MetaCore software pathway analysis showed 153 load-regulated molecular pathways in which significantly regulated genes were present. Particulary observed was an up-regulation of dentin matrix protein 1 (DMP-1, 2.18 fold) which plays a critical role for bone mineralization and strength; Wnt5a (2.19 fold) an osteoblast differentiation activator of Wnt signaling; and alpha-actinin (6.3 fold) which is involved in the cellular mechanoprotective response and raising the amount of alpha-actinin in the cytoskeleton drives to increase the whole cell resistance to deformation. Quantitative real-time polymerase chain reaction (PCR) results confirmed that the mRNA levels of Wnt5a and Asporin mRNA were indeed up-regulated.

- vii- In conclusion, this thesis has provided a method for isolation of RNA from trabecular osteocytes and using it for cDNA microarrays to reveal the mechanobiological effect of acute and chronic loading regimes in-vivo on global murine differential gene expression, including also analysis of signaling pathways. This analysis led to the identification of genes whose expression is regulated by mechanical loading thus pointing out potential molecular mechanisms involved in the osteogenic anabolic response to mechanical loading. This in turn paves the way to studying the role of genes in load-stimulated bone formation in corresponding genetically modified systems.

- viii- Zusammenfassung

Angesichts der vielen Knochenkrankheiten und –verletzungen, für welche angemessene Behandlungsmethoden erst noch entwickelt werden müssen, besteht die Notwendigkeit, sich mit neuen Ansätzen zu befassen, die die Knochenheilung stimulieren. Osteoporose ist eine Krankheit, welche durch eine übermässige Abnahme der Knochenmasse charakterisiert ist, was zu einer erhöhten Anfälligkeit für Brüche, Skelettdeformationen und - in schwerwiegenderen Fällen - zu Tod infolge Morbidität führen kann. Die Krankheit wurde sowohl auf genetische als auch auf altersbedingte Einflussfaktoren zurückgeführt. Verschiedenartige medikamentöse Behandlungen stehen zur Verfügung, welche Knochenverlust erwiesenermassen verzögern; allerdings ist noch keine Heilmethode greifbar. Um die Krankheit zu behandeln, versucht die medizinische Forschung, auf Gene abzuzielen, welche Osteoporose verursachen, indem sie die Maus als Tiermodell verwendet. Dank der kürzlichen Entschlüsselung des Erbgutes der Maus und infolge der grossen Homologie zwischen dem Chromosomensatz des Menschen und dem der Maus verkörpern Inzucht- Mäusestämme das ideale Modell für genetische Studien. Die Verwendung der Maus, um Gene zu identifizieren, welche in Prozessen des Knochenumbaus eine Rolle spielen, könnte zu Fortschritten im Verständnis der konkreten Regulierung bestimmter Knochenphänotpyen durch Gene und Proteine führen, wie beispielsweise im Falle der Knochenmineraldichte oder der Knochenstärke. Ein interessanter Phänotyp, der untersucht wird, ist das Knochenverhalten unter mechanischer Belastung beziehungsweise die mechanische Sensitivität. Mechanische Belastung ist wohl der wichtigste physiologische/umfeldbedingte Einzelfaktor, welcher die Knochenmasse und die Knochenform reguliert. Altersbedingter Knochenverlust und folgerichtig, Osteoporose, wurden zumindest teilweise einem Abbau der Muskelmasse/-funktion und der resultierenden Abnahme des mechanischen Gebrauchs des Skeletts zugeschrieben. Andererseits wurde gezeigt, dass mechanische Überbeanspruchung Knochenbildung fördert und einen Nettozuwachs an trabekulärer Knochenmasse verursacht, welche die wichtigste Baugruppe von lasttragenden Elementen im Skelett ausmacht. Allerdings weiss man nur wenig über die Mechanismen, welche innerhalb trabekulären Knochens an den lastbedingten anabolischen Wirkungen beteiligt sind. Dies ist vor allem darauf zurückzuführen, dass Studien lastbedingter molekularer Vorgänge im lebenden Organismus fehlen. Aufbauend auf Studien in unserer Gruppe, welche kürzlich die Auswirkungen mechanischer Belastung auf die trabukläre Knochenadaption im C56BL/6 Mausschwanz-Model untersucht hat, besteht die längerfristige Zielvorgabe dieser Doktorarbeit darin, die molekularen Mechanismen aufzuklären, welche mittels mechanischer

- ix- Belastung osteogenetische anabolische Auswirkungen hervorrufen und zudem, Gene und genetische Signalwege zu finden, welche durch mechanische Belastung reguliert werden. Der Chromosomensatz der Maus ist gut beschrieben und zugänglich für Kunstgriffe transgener und knockout Technologien. Das Verständnis molekularer Signalwege, welche den Knochenaufbau, angeregt durch mechanische Belastung, regelt, könnte die Möglichkeit bieten, die mechanische Sensitivität nachzuahmen oder zu erhöhen, indem pharmakologische und molekulare Mittel eingesetzt werden. Dies würde zur Entwicklung neuer Strategien für die Handhabung von Osteoporose und anderer Skelettmängel führen. Um die genetische Regulierung mechanischer Belastung zu untersuchen, wurde ein Methode ausserhalb des lebenden Organismus entwickelt, um repräsentative Proben intakter ribonukleinsäure (RNS) von Maus-Rückenwirbeln zu isolieren, welche selektiv von trabekulären Osteoblastzellen/Wandzellen und Osteozyten abgeleitet wurden, indem sequentielle Bindegewebe- Verdauung und -Zerstäbung angewandt wurden. Hochwertige Gesamt-RNS-Vorbereitungen wurden sofort nach der Zellseparierung isoliert, indem konventionelle Reagenzien und Protokolle verwendet bzw. befolgt wurden. Die Gesamtmenge an RNS, die von trabekulären Osteozyten eines einzigen Schwanzwirbelknochens isoliert wurden, war ausreichend für eine darauffolgende differenzielle Genexprimierungs-Analyse. Um eine einzelne und sich wiederholende, mittels mechanischer Belastung induzierte, differenzielle Genexpression zu untersuchen, wurde der fünfte Schwanzwirbelknochen von weiblichen C57BL/6 (B6)-Mäusen mechanisch stimuliert. Die Stimulierung wurde unter Verwendung von einzelnen respektive sich periodisch wiederholenden Lastdosierungen erreicht, wobei letztere aus 3000 Zyklen bestanden, die bei einer Frequenz von 10 Hz und einer Amplitude von 0N und 8N via zwei Stiften in den benachbarten Rückwirbeln (C4 und C6) übertragen wurden. Die Mäuse wurden sechs Stunden nach der letzten mechanischen Belastung getötet, um danach hochwertige Gesamt-RNS-Vorbereitungen für Genexprimierungs-Arrays zu untersuchen, indem Affymethrix Mouse Genome chips verwendet wurden. Differenzielle Genexprimierungs-Analyse von einzeln applizierten Lastereignissen legte insgesamt 359 sifnifikant regulierte Gene (P<0.05) offen, 301 davon hochreguliert und 58 runterreguliert. Zudem deutete eine mittels dem Computerprogramm GeneGo MetaCore durchgeführter funktioneller Genomik-Analyse von akuten Lastereignissen auf 65 lastregulierte molekulare Signalwege hin, wobei signifikant regulierte Testergebnisse verzeichnet werden konnten. Namentlich wurden eine Hochregulation des Insulin- Wachstumsfaktors 1 (IGF-1, 2.2-fach) und flügellosartige MMTV Integrationsseite-Familie, Mitglied 5a (Wnt5a, 3.4-fach) nachgewiesen, von welchen vermutet wird, dass sie Auslöser der

- x- Osteoblast-Differenzierung sind. Demgegenüber wurde eine Herunterregulation des WNT- Hemmfaktors 1-Gen (WIF-1, 1.8-fach) ermittelt, welcher ein Hemmer des WNT/beta-Catenin- Signalwegs ist und die Aktivierung der Osteoblast-Differenzierung unterdrückt. Differenzielle Genexprimierungs-Analyse von sich wiederholenden Lastereignissen (drei Mal pro Woche ein einzelnes Lastereignis, insgesamt über vier Wochen) hat 1585 signifikant hochregulierte Testergenisse geliefert, wobei 860 hochregulierte und 725 herunterregulierte Gene. Ferner wurden mittels einer mit dem Computerprogramm MetaCore durchgeführten Signalweg-Analyse 153 lastregulierte molekulare Signalwege ermittelt, in welchen signifikant regulierte Gene gegenwärtig sind. Im Einzelnen wurde eine Hochregulation des Dentin-Gewebeproteins 1 (DMP-1, 2.18-fach) beobachtet, welches eine entscheidende Rolle spielt für die Knochenmineralisierung und -stärke; Wnt5a (2.19-fach), ein Osteoblast-Differenzierungs-Auslöser der Wnt-Signalübertragung; und alpha-Actinin (6.3-fach), welches in der zellulären Antwort zum mechanischen Schutz und Aufhebung des Betrags des alphas-Actinin in den Cytoskeleton-Laufwerken beteiligt wird, um den ganzen Zellwiderstand gegen die Deformierung zu vergrössern. Resultate quantitativer Echtzeit- Polymerase-Kettenreaktion (PCR) bekräftigte, dass die mRNS-Niveaux von Wnt5a und Asporin in der Tat hochreguliert waren. Zusammenfassend hat diese Doktorarbeit eine Methode für die RNS-Isolierung aus trabekulären Osteozyten vorgelegt, welche für cDNA-Microarrays angewendet wurde, um die mechano- biologischen Auswirkungen akuter und chronischer Last-Regimes auf globale differentielle Genexprimierung in der Maus in vivo offenzulegen, wobei auch die Analysie von Signalwegen berücksichtigt wurde. Diese Untersuchungen hat zur Ermittlung von Genen geführt, deren Exprimierung durch mechanische Belastung reguliert ist und somit potentielle molekulare Mechnismen aufzeigt, welche an osteogenetischen anabolischen Reaktionen auf mechnische Belastung beteilligt sind. Dies wiederum ebnet den Weg, um die Funktionen von Genen zu studieren, welche sie innerhalb genetisch modifizierter System für lastinduzierte Knochenbildung ausüben.

- xi- Chapter 1: Introduction Chapter 1

- 12- Chapter 1: Introduction Introduction

The Institute for Biomechanics aims to provide a bridge between biologists, who bring molecular and cellular components to engineering, and engineers, who bring the methods of measurement, analysis, synthesis, and control to molecular and cell biology. Biomechanics develops, refines, and uses bioengineering tools and concepts to explore and understand living systems on the molecular, cellular and organic levels. Bone provides life-essential functions as the framework of the body, protecting inner organs from injury, and representing a storehouse for vital minerals. Nevertheless, because bone is subject to pathologies or injuries, its ability to exert its essential functions may be lost. For example, osteoporosis is a disease characterized by an excessive decrease in bone mass which can lead to an increased susceptibility to fractures, skeletal deformation and, in more severe cases, death owing to morbidity. The disease has been attributed to both genetic and age-related factors. Appart from the obvious costs on health, osteoporosis is a global problem and carries with it significant social and economic costs. This is illustrated by the IOF audit report “Call to Action” published in 2001, which claims that osteoporosis costs national treasuries in the EU over 4.8 billion Euro annually in hospital healthcare alone. Various medications are available which have been shown to delay bone loss, however no cure is yet achievable (1). The concept of bone mass homeostasis maintained by mechanical loads is widely accepted and supported by a substantial body of experimental evidence (2). Medical research is now attempting to target genes which define osteoporosis using the mouse as a model system for human diseases. Owing to the recent deciphering of the mouse genome and the high homology that exists between the human and mouse genomes (3), inbred strains of mice represent ideal models for genetic studies. Using the mouse to identify genes implicated in the bone remodeling process could lead to advances in understanding that enable the precise regulation of the genes and proteins responsible for particular bone phenotypes, i.e. bone mineral density or bone strength. One interesting phenotype under investigation is the response of bone to mechanical loading or its ‘mechano-sensitivity’. Bone is a specialized connective tissue with basically two functions, both of which are related to its unique characteristic as a calcified extracellular matrix. The first function is to carry heavy mechanical loads, derived from weight bearing or from muscle contractions. The second function is to serve as a reservoir of ions such as calcium, phosphate and magnesium, whereby bone tissue helps to maintain the homeostasis of these ions in the blood. Bone is a living, continuously self- renewing tissue. At first sight, the most important cell types involved in the formation, modeling

- 13- Chapter 1: Introduction and remodeling of bone are the osteoblasts or bone-forming cells and the osteoclasts, the bone- resorbing cells. The most abundant cell type in mature bone is, however, the osteocyte. There are approximately 10 times as many osteocytes as osteoblasts in normal human bone (4). Osteocytes have a particular location in bone. During bone formation some osteoblasts are left behind while the bone formation front moves on together with the other retracting osteoblasts. The encapsulated osteoblasts differentiate into osteocytes. They lose a large part of their cell organelles but gain long slender cell processes by which the cells remain in contact with earlier incorporated osteocytes and with osteoblasts lining the bone surface (22). Despite the relative abundance of osteocytes in bone tissue, they have not yet been shown to have an unequivocal function. Their location in bone and their organization in a syncytium with two extensive communication systems, one intracellular (osteocyte-gap junction-osteocyte) and another extracellular (lacuna-canaliculus-lacuna) suggest at least two possible ways in which osteocytes may function: 1) to ensure communication between sites deep in the bone and the extra-osseous world and 2) to create an enormous increase in mineral surface exposed to extracellular fluid and cellular activity. These considerations have led to the formulation of the following hypotheses about the function of the osteocyte.

1.1. Hypotheses

The primary function of the skeleton is to bear mechanical loads, a principal reason for the existence of a hard, mineralized extracellular matrix. It has long been recognized that the amount of mechanical loading to which a piece of bone is exposed and the geometry and mass of that bone are related. Living bone is continually undergoing processes of remodeling; this allows a continuous fine tuning of the amount and spatial organization of the tissue, to provide maximal strength with a minimum of bone mass. This process is called functional adaptation and was originally described as Wolff's law about 100 years ago (5). Although it is generally considered that functional adaptation is achieved by the concerted action of osteoblasts and osteoclasts, the mechanism by which these cells are instructed for such a task remains obscure. To bring about meaningful change in existing bone tissue, osteoblasts and osteoclasts must be informed about local needs for tissue increase or reduction; these in turn depend on mechanical overuse or underuse. Both osteoblasts and osteoclasts act at the surface of bone tissue, while mechanical loads produce displacements, or strains, throughout the bone. Thus, aberrant strain would best be detected by living elements dispersed throughout the matrix. Osteocytes are the only cells that can fulfill this demand. Sensor cells that detect loading deviations need not also be actor cells that carry out adaptation, as long as sensors

- 14- Chapter 1: Introduction and actors can communicate with one another. In this respect the organization in bone of a cellular network, where osteocytes, embedded in the matrix, are connected via cell processes and gap junctions to osteoblasts on the surface of that matrix, assumes significance. Mechano-transduction in bone involves a number of steps. First, the mechanical load or stress must be transduced into a physical signal that is sensed by the bone cells. This process is called mechanical coupling. It is not known which physical signal resulting from stress performs this function in bone. Mechanical loads produce deformation or strain gradients within the bone tissue; these in turn cause fluid to flow through the canalicular network (6, 7). Strain resulting from stress, or flow resulting from strain or both might activate bone cells. In the second step of mechano- transduction, the physical signal is translated by the cell into a biochemical signal. This may be called biochemical coupling. Many studies have shown that osteocytes in culture react to physical stress with an enhanced production of prostaglandins, primarily prostaglandin E2 (PGE2). Second messengers such as cyclic adenosine monophosphates (cAMPs) are also produced (8). In the third step, the biochemical signal must be communicated to effector cells, i.e., the osteoblasts and osteoclasts, which react by augmenting or reducing the amount of bone matrix at a specific site. Is there a role for osteocytes in mechano-transduction? Many authors in recent papers speculate that this is so, and experimental studies by Lanyon (9) have provided some evidence for this hypothesis, having shown that osteocytes change metabolic activity when subjected to strain. Recently, a number of experiments in which isolated osteocytes were subjected to two types of mechanical stress in vitro have been performed (10). The results suggest that osteocytes are indeed very sensitive to stress, responding by enhanced production of prostaglandins and other factors. Thus, our hypotheses are: 1) the anabolic effect of mechanical load in trabecular bone is mediated by osteocytes and involvs a change in gene expression; 2) some of the differentially regulated genes affect bone remodeling. In an effort to continue the search for a cure, this project aims to elucidate the genetic and biochemical factors of bone formation in response to mechanical loading. It has been demonstrated in humans and experimental animals that cyclic overloading results in increased bone formation and a net gain in lamellar bone mass, both cortical and trabecular (11-14). However, very little is known about the molecular mechanisms involved in the load stimulated bone formation, especially in trabecular bone which occupies the critical load-bearing sites of the skeleton (vertebrae, hip, distal radius).

- 15- Chapter 1: Introduction Recently, the establishment in our group by Webster et al. 2008 of a mouse model for load-induced anabolic activity in trabecular bone has facilitated molecular approaches in this field because mice have a well-characterized genome accessible to manipulations by transgenic and knockout technologies. Building on this model, this study allows the isolation of a well-defined osteocyte population from cancellous bone for detailed molecular characterization. To this end, we devised a strategy designed to elucidate load-induced molecular changes in trabecular bone. This strategy is based on the cyclic compression of individual mouse caudal vertebrae followed by expression analysis in total RNA isolates from selective osteocytic enriched cell population. A successful implementation of this strategy is expected to provide a method and baseline for further studies aimed at elucidating the molecular mechanisms involved in the trabecular adaptation to mechanical loads. We expect our strategy to yield meaningful advances in the understanding of bone adaptation to mechanical stimuli, thus uncovering skeletal mechanisms involved in the skeletal load-bearing function and their role in osteoporosis.

1.2. Specific Aims

1.2.1. Developing a method of isolating intact mRNA from a well-defined trabecular osteocytic enriched cell population of murine caudal vertebra in the C57BL/6 mouse strain. Hypothesis to be tested: Assess the feasibility of obtaining a sufficient amount of high quality total RNA from trabecular osteocytes derived from single caudal vertebra for microarray analysis. Well-defined trabecular osteocytic population shall be targeted by separation of trabecular bone from caudal vertebral body, collagenase digestions of medullary soft tissue and pulverization of remaining trabeculae. Enzymatic “stripping” of cells will be confirmed histologically following each digestion step. Total RNA from collected bone cell fractions will be subjected to gene expression profiling by reverse transcripted-polymerase chain reaction (RT-PCR).

1.2.2. Identify load-induced differentially regulated gene expression using trabecular osteocytic RNA isolated from caudal vertebra of C57BL/6 mice following the administration of a single and repetitive dose of cyclic mechanical load. Hypothesis to be tested: A single and repetitive dose of cyclic mechanical loading induces changes in the expression of osteocytic genes. A target caudal vertebra will be cyclically compressed via pins inserted into adjacent vertebrae. Total RNA from trabecular osteocytes will be isolated by the protocol established in Specific Aim

- 16- Chapter 1: Introduction 1.2.1 and their mRNA amplified and analyzed for differential gene expression using cDNA microarrays. The differential gene expression will be further analyzed to compare transient and sustained gene expression following a single loading regimen and a repetitive loading regimen respectively. Genes and molecular pathways will be identified using functional genomics approaches including advanced bioinformatics tools in close collaboration with the Functional Genomics Center Zurich. Differentially regulated gene expression of interest will be confirmed by real-time RT-PCR.

1.3. Outline of the thesis

This thesis is composed of 5 chapters. In order to put the novelty of the proposed research project into context, chapter 2 will introduce the bone anatomy and physiology, in particular bone remodeling and the skeletal role of osteocytes as mechanosensors. Relevant aspects of bone cell and molecular biology relevant loading will be also covered in this chapter. Further, strategies for finding the genetic determinants related to load-induced bone adaptation in treatment of osteoporosis will be discussed. This latter point will emphasize the significance of the mouse as a model for human diseases, demonstrating the potential of this project to yield meaningful advances in understanding bone adaptation to mechanical stimuli and thus uncovering the skeletal mechanisms involved in the load-bearing function and their role in osteoporosis. Chapter 3 presents the development of a robust method for isolating osteocytic RNA from a well-defined trabecular osteocytic population of a single caudal vertebra to study load-induced molecular events in osteocytes. This chapter is organized into sections which describe a specific strategy and steps of the established protocol. Firstly, the developed method isolates intact mRNA from well-defined trabecular osteocytes using minimal numbers of the animals in the study. The first step in this process was to physically separate the trabecular bone from the medullar cavity of the target caudal vertebra without contamination from cortical bone. This initial step is the most important in determining the yield of the final amount of mRNA. Additional sections present sequential enzymatic digestion for obtaining a well-defined selective osteoblast/lining cell population, followed by pulverization of denuded trabeculae and RNA extraction of intact osteocytic RNA. The accuracy of enzymatic removal of the appropriate cells was confirmed histologically following each digestion step. Extracted total RNA was then analyzed on quality and quantity with subsequent gene expression profiling of different bone cell isolates. A minimum of 0.5 ng of total RNA was required for a single cDNA microarray run. Before attempting to identify load induced gene expression, this

- 17- Chapter 1: Introduction protocol was evaluated and optimized by applying it to groups of non-loaded mice. Chapter 4 presents the results of load-induced changes in gene expression in trabecular osteocytes, by the method developed in chapter 3 and the comparison of differentially regulated genes between single and repetitive mechanical loads. Functional genomics for the identification of load-regulated molecular pathways was then outlined and the load-regulated expression of individual genes after single loading was confirmed by quantitative polymerase chain reaction. Finally, the integration in Chapter 5 brings together the results and discusses the benefits and limitations of the presented work, outlining future steps to further advance this field of research.

- 18- Chapter 1: Introduction References

1. Cosman F 2005 The prevention and treatment of osteoporosis: a review. Medscape General Medicine 7(2):73.

2. Rodan GA 1997 Bone mass homeostasis and bisphosphonate action. Bone 20(1):1-4.

3. Waterston RH, Lindblad-Toh K, Birney E, Rogers J, Abril JF, Agarwal P, Agarwala R, Ainscough R, Alexandersson M, An P, Antonarakis SE, Attwood J, Baertsch R, Bailey J, Barlow K, Beck S, Berry E, Birren B, Bloom T, Bork P, Botcherby M, Bray N, Brent MR, Brown DG, Brown SD, Bult C, Burton J, Butler J, Campbell RD, Carninci P, Cawley S, Chiaromonte F, Chinwalla AT, Church DM, Clamp M, Clee C, Collins FS, Cook LL, Copley RR, Coulson A, Couronne O, Cuff J, Curwen V, Cutts T, Daly M, David R, Davies J, Delehaunty KD, Deri J, Dermitzakis ET, Dewey C, Dickens NJ, Diekhans M, Dodge S, Dubchak I, Dunn DM, Eddy SR, Elnitski L, Emes RD, Eswara P, Eyras E, Felsenfeld A, Fewell GA, Flicek P, Foley K, Frankel WN, Fulton LA, Fulton RS, Furey TS, Gage D, Gibbs RA, Glusman G, Gnerre S, Goldman N, Goodstadt L, Grafham D, Graves TA, Green ED, Gregory S, Guigo R, Guyer M, Hardison RC, Haussler D, Hayashizaki Y, Hillier LW, Hinrichs A, Hlavina W, Holzer T, Hsu F, Hua A, Hubbard T, Hunt A, Jackson I, Jaffe DB, Johnson LS, Jones M, Jones TA, Joy A, Kamal M, Karlsson EK, et al. 2002 Initial sequencing and comparative analysis of the mouse genome. Nature 420(6915):520-62.

4. Parfitt AM. 1977 The cellular basis of bone turnover and bone loss: a rebuttal of the osteocytic resorption - bone flow theory. Clin Orthop Relat Res. 127:236–47.

5. Wolff J. 1892 Das Gesetz der Transformation der Knochen. Hirschwald Verlag, Berlin.

6. Piekarski K. and Munro M. 1977 Transport mechanism operating between blood supply and osteocytes in long bones, Nature 269 pp. 80–82.

7. Weinbaum SA, Cowin SC, Zeng YA. 1994 A model for the excitation of osteocytes by mechanical loadinginduced bone fluid shear stresses. J Biomech Mar; 27(3):339–60.

8. Burger E. and Veldhuijzen JP. 1993 Influence of mechanical factors on bone formation, resorption, and growth in vitro. In: B.K. Hall, Editor, Bone Vol. 7, CRC Press, Boca Raton, Florida pp. 37–56.

9. Lanyon LE. 1993 Osteocytes, strain detection, bone modeling and remodeling. Calcif Tissue Int 53:S102-106, discussion S106-107.

10. Klein-Nulend J, van der Plas A, Semeins CM, Ajubi NE, Frangos JA, Nijweide PJ, Burger EH 1995 Sensitivity of osteocytes to biomechanical stress in vitro. Faseb J 9(5):441-5.

11. Evans, WJ 1998 Exercise and nutritional needs of elderly people: effects on muscle and bone. Gerodontology 15(1):15-24.

12. Duncan RL, Turner CH 1995 Mechanotransduction and the functional response of bone to mechanical strain. Calcif Tissue Int 57(5):344-58.

- 19- Chapter 1: Introduction 13. Biewener AA, Fazzalari NL, Konieczynski DD, Baudinette RV 1996 Adaptive changes in trabecular architecture in relation to functional strain patterns and disuse. Bone 19(1):1-8.

14. Layne JE, Nelson ME 1999 The effects of progressive resistance training on bone density: a review. Med Sci Sports Exerc 31(1):25-30.

15. Lean JM, Mackay AG, Chow JW, Chambers TJ 1996 Osteocytic expression of mRNA for c-fos and IGF-I: an immediate early gene response to an osteogenic stimulus. Am J Physiol 270(6 Pt 1):E937-45.

16. Hillam RA, Skerry TM 1995 Inhibition of bone resorption and stimulation of formation by mechanical loading of the modeling rat ulna in vivo. J Bone Miner Res 10(5):683-9.

17. Forwood MR, Turner CH 1995 Skeletal adaptations to mechanical usage: results from tibial loading studies in rats. Bone 17(4 Suppl):197S-205S.

18. Webster D, Morley Pl, van lemthe GH, and Müller R. 2008 A novel in vivo mouse model for mechanically stimulated bone adaptation – a combined experimental and computational validation study. Comput Methods Biomech Biomed Engin 11(5):435-41.

19. Webster D, Wasserman E, Weber F, Bab I and Müller R. 2008 Load induced changes in trabecular and cortical bone are dose dependent in both C57Bl/6 and C3H/Hej mice. Abstracts 30th Annual Meeting ASBMR, Montreal, Canada, J. Bone Miner. Res. 23:S131.

20. Xing W, Baylink D, Kesavan C, Hu Y, Kapoor S, Chadwick RB, Mohan S 2005 Global gene expression analysis in the bones reveals involvement of several novel genes and pathways in mediating an anabolic response of mechanical loading in mice. J Cell Biochem 96(5):1049-60.

21. Lau KH, Kapur S, Kesavan C, Baylink DJ 2006 Up-regulation of the Wnt, estrogen receptor, insulin-like growth factor-I, and bone morphogenetic protein pathways in C57BL/6J osteoblasts as opposed to C3H/HeJ osteoblasts in part contributes to the differential anabolic response to fluid shear. J Biol Chem 281(14):9576-88.

22. Palumbo C, Palazzini S, Marotti G. 1990 Morphological study of intercellular junctions during osteocyte differentiation. Bone 11(6):401–406.

23. Turner CH, Hsieh YF, Müller R, Bouxsein ML, Baylink DJ, Rosen CJ, Grynpas MD, Donahue LR, Beamer WG 2000 Genetic regulation of cortical and trabecular bone strength and microstructure in inbred strains of mice. J Bone Miner Res 15(6):1126-31.

24. Robling AG, Turner CH 2002 Mechanotransduction in bone: genetic effects on mechanosensitivity in mice. Bone 31(5):562-9.

25. Han Y, Cowin SC, Schaffler MB, Weinbaum S 2004 Mechanotransduction and strain amplification in osteocyte cell processes. Proc Natl Acad Sci USA 101(47):16689-94.

- 20- Chapter 2: Background Chapter 2

- 21- Chapter 2: Background Background

2.1. Bone Anatomy

Bone is a specialized connective tissue that together with cartilage constitutes the skeletal system. These tissues provide the organism with a rigid structure which, with muscle, facilitates stable locomotion. Like reinforced concrete, the bone matrix is predominantly a mixture of tough fibers (made of type I collagen) which resist pulling forces, and solid particles (calcium phosphate as hydroxyapatite crystals) which resist compression. For all its rigidity, bone is by no means an inert substance but a living organ. Running throughout the hard extra cellular matrix are channels and cavities occupied by living cells which account for about 15 % of the weight of compact bone. These cells are engaged in the life long process of remodeling the bone: one class of cells (osteoclasts) demolishes old bone matrix while another (osteoblasts) deposits new bone matrix in the interior of the bone (1). All bony skeletal structures have an external shell made of a dense layer of calcified matrix, referred to as the cortex or compact bone (2). Their internal space is partially filled with a boney mesh, referred to as cancellous bone, spongy bone or trabecular bone (Fig. 2.1). The spaces enclosed by these trabeculae are filled with hematopoietic bone marrow which also fills the trabecular-free medullary cavity. The periosteal, endosteal and trabecular surfaces are lined with fibroblastic and osteogenic cells organized in layers, which comprise the periosteum and endosteum. Cortical bone, which comprises 80% of the skeleton, is dense and compact, has a low turnover rate and a high resistance to bending and torsion. The function of the cortical bone is to provide mechanical strength and protection, but it can also participate in metabolic responses, particularly when there is severe or prolonged mineral deficit. Trabecular bone represents 20% of the skeletal mass but 80% of the bone surface is found inside the long bones throughout the bodies of the vertebrae, in the femoral neck and in the inner portions of the pelvis and other flat bones. Trabecular bone is less dense, more elastic, and has a high turnover rate. Trabecular bone contributes to mechanical support, particularly in bones such as the vertebrae and femoral neck, and provides the initial supplies of mineral in acute deficiency states.

- 22- Chapter 2: Background

1 mm

Endosteum Periosteum

Marrow cavity

Cortical Bone Trabecular Bone

Fig. 2.1: The anatomical structure of caudal vertebra (Bab et al. Micro-tomographic Atlas of the Mouse Skeleton. 2007).

2.1.1. Bone Matrix

Bone matrix mainly consists of type I collagen fibers (consisting of two α1 chains and one α2 chain) and non-collagenous proteins. Within lamellar bone, the fibers form arches that allow the highest density of collagen per unit volume of tissue. The lamellae can run parallel to each other (trabecular bone and periosteum), or be concentric surrounding a channel centered on a blood vessel (cortical bone Haversian system). Crystals of hydroxyapatite [Ca3(PO4)2(OH)2] are found in, around and between the collagen fibers and tend to be oriented in the same direction as the collagen fibers. The role of numerous noncollagenous proteins present in the bone matrix has not been fully explained. The major noncollagenous protein produced is osteocalcin (bone Gla protein), whose function in bone is still unknown, osteonectin (phosphoprotein), bone sialoprotein (glycoprotein), osteopontin and bone morphogenetic proteins which also may play an important role in bone mineralization (4-

- 23- Chapter 2: Background 6). Biglycan, a proteoglycan, is expressed in the bone matrix and probably positively regulates bone formation (7). There are four types of bone cells:

1. Osteoblast - mononucleated cell responsible for bone formation (Fig. 2.2). Osteoblasts do not function individually but are found in sheaths along the bone surface on the layer of bone matrix that they are producing. They originate from multipotent stromal stem cells, which have the capacity to differentiate into osteoblasts, adipocytes, chondrocytes, myoblasts, or fibroblasts (8). Gene deletion studies have shown that absence of runtrelated transcription factor 2 (Runx2) or of a downstream factor, osterix, is critical for osteoblast differentiation (9). Toward the end of the matrix-secreting period, 15% of mature osteoblasts are entrapped in the new bone matrix and differentiate into osteocytes, whereas some cells remain on the bone surface, becoming flat lining cells. Bone formation occurs in three successive phases: the production and the maturation of osteoid matrix, followed by mineralization of the matrix. In normal adult bone, these processes occur at the same rate so that the balance between matrix production and mineralization is equal. Initially, osteoblasts produce osteoid by rapidly depositing collagen. This is followed by an increase in the mineralization rate to equal that of collagen synthesis. In the final stage the rate of collagen synthesis decreases and mineralization continues until the osteoid becomes fully mineralized. Osteoblasts produce a range of growth factors under a variety of stimuli including the insulin-like growth factors (IGF), platelet-derived growth factor (PDGF), basic fibroblast growth factor (bFGF), transforming growth factor-beta (TGF- β) and the bone morphogenetic proteins (BMP), (10-14). These factors regulate bone remodeling in an autocrine/paracrine manner by activating receptors found on osteoblasts. In turn, the activity of these factors is controlled by systemic endocrine factors such as the gonadal, parathyroid and pituitary hormones and by the autonomic nervous system (15- 23).

2. Osteoclast – a giant multinucleated cell up to 100 mm in diameter derives from hematopoietic cells of the mononuclear lineage (26) and is the bone lining cell responsible for bone resorption. It is usually found in contact with a calcified bone surface and within a lacuna (Howship’s lacunae) as a result of its own resorptive activity (Fig. 2.3). Osteoclasts have abundant Golgi complexes, mitochondria, and transport vesicles loaded with lysosomal enzymes. They present deep foldings of the plasma membrane in the area facing the bone matrix (called ruffled border) and the surrounding zone of attachment (called sealing zone).

- 24- Chapter 2: Background Lysosomal enzymes such as tartrate-resistant acid phosphatase and cathepsin K are actively synthesized by the osteoclast and are secreted via the ruffled border into the bone-resorbing compartment (27). The process of the osteoclast attachment to the bone surface involves the binding of integrins expressed in osteoclasts with specific amino acid sequences within proteins at the surface of the bone matrix (28). After osteoclast adhesion to the bone matrix, αvβ3 integrin binding activates cytoskeletal reorganization within the osteoclast (29). Attachment usually occurs via dynamic structures called podosomes. Through their continual assembly and disassembly they allow osteoclast movement across the bone surface during which bone resorption proceeds. Integrin signaling and subsequent podosome formation is dependent on a number of adhesion kinases including the proto-oncogene src (30). Osteoclasts resorb bone by acidification, disintegration of the hydroxyapatite crystals and proteolysis of the bone matrix and of the hydroxyapatite crystals encapsulated within the sealing zone. Osteoclast function is regulated both by locally acting cytokines, mainly macrophage-colony stimulating factor (M-CSF), receptor activator of NF-kappa B ligand (RANKL) and osteoprotegerin (OPG) and by systemic hormones (31-39).

3. Bone-lining cells – essentially inactive osteoblasts on the surface of most bones in an adult, which are responsible for calcium and phosphate exchange in the bone (Fig. 2.3).

Fig. 2.2: Deposition of bone Osteogenic cell matrix by osteoblasts: (osteoblast precursor) osteoblasts lining the surface of bone secrete the organic Osteoblast matrix of bone (osteoid) and Osteoid (uncalcified bone are converted into osteocytes matrix) as they become embedded in the matrix. The osteoblasts Calcified bone matrix themselves are thought to derive from osteogenic stem cells that are closely related to Cell process in canaliculus fibroblasts (Alberts et al. Osteocyte Molecular biology of the cell. 4 ed. 2002).

Bone is a dynamic tissue constantly being reshaped by osteoblasts, which build bone and osteoclasts, which resorb bone (Fig. 2.3).

- 25- Chapter 2: Background Fig. 2.3: The remodeling of Quiescent osteoblast compact bone: osteoclasts tunnel (bone lining cell) through old bone, while osteoblasts form new bone Small blood vessel (Alberts et al. Molecular biology of the cell. 4 ed. 2002). Endothelial cell New bone

Fibroblast New bone matrix not yet calcified

Osteocyte Osteoclast tunneling Osteoblast about to lay down through old bone new bone

4. Osteocyte - a mature osteoblast surrounded by bone matrix, the most abundant cells found in bone (Fig. 2.2). Even though the metabolic activity of the osteoblast decrease once it is fully encased in lacunae surrounded by bone matrix, these cells still produce matrix and regulatory proteins (177). Osteocytes have numerous long cell processes rich in microfilaments that are organized during the formation of the matrix and before its calcification. They form a network of narrow canaliculi permeating the entire bone matrix. Osteocyte morphology varies according to cell age. A young osteocyte has most of the structural characteristics of the osteoblast but a decreased cell volume. An older osteocyte, located deeper within the calcified bone, exhibits a further decrease in cell volume and an accumulation of glycogen in the cytoplasm. The osteocytes are finally phagocytosed and digested during osteoclastic bone resorption (24). Despite the complex organization of the osteocytic network, the exact function of these cells remains purely understood. Recent evidence suggests that osteocytes produce sclerostin, a protein that inhibits bone formation and whose expression is decreased by mechanical stimuli (25).

2.2. Osteocytes as mechanosensors/endocrine and paracrine function

Osteocytes, composing over 90–95% of all bone cells in the adult animal (121), are defined as cells embedded in the mineralized bone matrix, but clear functions have not yet been ascribed to these cells, unlike to osteoblasts and osteoclasts. Osteocytes are regularly dispersed throughout the mineralized matrix within ‘caves’ called lacunae, connected to each other and cells on the bone surface through slender, cytoplasmic processes or dendrites passing through the bone in thin

- 26- Chapter 2: Background ‘tunnels’ (100–300 nm) called canaliculi (Fig. 2.4). Not only do these cells communicate with each other and with cells on the bone surface, but their dendritic processes are also in contact with the bone marrow (122), implying that osteocytes can communicate with marrow resident cells. One means for communication with other cell types is through gap junctions, and another is through release of signaling molecules into the bone fluid that flows through the lacuno-canalicular system. The most popular theory regarding the major function of osteocytes is that they translate mechanical strain into biochemical signals between osteocytes and to cells on the bone surface to affect (re)modeling (123), but this yet remains to be definitively proven. Recent data suggest additional important functions for osteocytes, such as the regulation of mineral metabolism (124) and the alteration of the properties of their surrounding matrix (125).

Fig. 2.4. Scanning electron micrograph of isolated osteocytes. (C) Two osteocytes have made contact with each other via their cell processes after 24 h of culture. (D) Extensive network of flattened osteocytes with many branched cell processes after 48 h of culture. Scale bar = 10 μm. (Reproduced from J. Bone Miner. Res. 1992; 7, 389–396 with permission of the American Society for Bone and Mineral Research.)

2.2.1. Osteocytes as mechanosensors directing bone formation and/or resorption

A known key regulator of osteoblast and osteoclast activity in bone is mechanical strain. The skeleton is able to continually adapt to mechanical loading by adding new bone to withstand increased amounts of loading, and by removing bone in response to unloading or disuse (reviewed in (126,127)). Galileo, in 1638, is documented as first suggesting that the shape of bones is related to loading. Julius Wolff, in 1892, more eloquently proposed that bone accommodates or responds to strain. The cells of bone with the potential for sensing mechanical strain and translating these forces into biochemical signals include bone lining cells, osteoblasts, and osteocytes. Of these, the osteocytes, with their sheer numbers and distribution throughout the bone matrix and their high

- 27- Chapter 2: Background degree of interconnectivity, are thought to be the major cell type responsible for sensing mechanical strain and translating that strain according to the intensity of the strain signals (123). Various studies have demonstrated load-related responses in osteocytes in vivo, supporting their proposed role as mechanotransducers in bone. Within a few minutes of loading, glucose-6- phosphate dehydrogenase, a marker of cell metabolism, is increased in osteocytes and lining cells (128). By 2 hours, c-fos mRNA is evident in osteocytes and by four hours, transforming growth factor-β and insulin-like growth factor-1 mRNAs are increased (129). Additional osteocyte selective markers, such as E11/gp38, dentin matrix protein 1 (DMP1), MEPE, and sclerostin, are also regulated by mechanical loading. The DMP-1, is activated in a few hours in response to mechanical loading in osteocytes in the tooth movement model (130) and in the mouse ulna loading model of bone formation (131). E11/gp38, a membrane protein that is osteocyte-selective and thought to play a role in dendrite elongation, is also activated within 4 hours after mechanical load, not only in cells near the bone surface, but also in deeply embedded osteocytes (132). As detailed below, the osteocyte specific marker sclerostin, the protein product of the SOST gene, is decreased in response to anabolic loading (133). Anabolic signals that are released within seconds after loading in osteocytes include nitric oxide (NO), prostaglandins, and other small molecules such as ATP. NO, a short-lived free radical that inhibits resorption and promotes bone formation is generated within seconds in both osteoblasts and osteocytes in response to mechanical strain (134). Primary osteocytes and primary calvarial bone cells have also been shown to release prostaglandins in response to fluid flow treatment, and a number of studies have suggested that osteocytes are the primary source of these load-induced prostaglandins (135). In vivo studies have shown that new bone formation induced by loading can be blocked by the prostaglandin inhibitor, indomethacin (136), and agonists of the prostaglandin receptors have been shown to increase new bone formation (137). Another anabolic pathway that appears to be activated rapidly in osteocytes within one hour in response to load is the canonical Wnt/β-catenin pathway. Johnson and colleagues, discoverers of the high bone mass (HBM) gene, a mutated low-density lipoprotein receptor-related protein 5 gene (LRP5) encoding the LRP5 receptor, hypothesized as early as 2002 that LRP5 is a major player in the way that bone cells respond to mechanical load (138). They reasoned that the HBM mutation results in a skeleton that is over-adapted in relation to the actual loads being applied, but yet is in homeostatic equilibrium. They found that wild-type bone experienced 40% greater strain than HBM bone with the same load (139). Based on these observations in humans and mice, they hypothesized

- 28- Chapter 2: Background that the set-point for load responsiveness was lower in the HBM skeleton. Loss of function mutations in LRP5 result in low bone mass, and mice with mutations in LRP5 do not respond to mechanical load (140), again supporting the notion that LRP5 is involved in mechanotransduction. At the most recent annual meeting of the ASBMR, Robling et al. showed that sclerostin, an inhibitor of the Wnt pathway that binds to LRP5 and that is produced exclusively by mature osteocytes, decreases 24 hours after loading (133). These investigators proposed that Wnt/β-catenin is the initiator and SOST/sclerostin is the inhibitor of load-induced new bone formation. Also at this meeting, Kamel et al. showed that prostaglandin released by bone cells in response to fluid flow can activate the Wnt/β-catenin pathway independent of LRP5 (141). These investigators suggested that prostaglandin can bypass the inhibitory effects of sclerostin present in the bone matrix. Osteocytes may also send signals for bone resorption. Isolated avian osteocytes have been shown to support osteoclast formation and activation (142), as has the osteocyte-like cell line, MLO-Y4. However, unlike any previously reported stromal cell lines, MLO-Y4 cells did so in the absence of any osteotropic factors (143). These cells express RANKL along their dendritic processes and secrete large amounts of macrophage colony-stimulating factor, both essential for osteoclast formation. Expression of RANKL along osteocyte dendritic processes, and the capacity of osteocyte dendritic processes to extend into the marrow space (122) provide a potential means for osteocytes within bone to interact and stimulate osteoclast precursors at the bone surface. Another means by which osteocytes can support osteoclast activation and formation is through apoptosis. Osteocyte apoptosis occurs at sites of microdamage, where the dying osteocyte may send signals to osteoclasts for targeted removal of bone (144). Investigators found that Bax (apoptotic biomarker) was elevated in osteocytes immediately at the microcrack locus, whereas Bcl-2 (anti-apoptotic biomarker) was expressed 1–2 mm from the microcrack, suggesting that damaged osteocytes send signals of resorption, whereas those osteocytes that do not undergo apoptosis are prevented from doing so by active protective mechanisms. It is still unclear if signals of resorption sent by dying osteocytes are the same or different from those sent by viable osteocytes. The parameters for inducing bone formation or bone resorption in vivo are fairly well-known and well-characterized. Bone mass is influenced by peak applied strain (145), and bone formation rate is related to loading rate (146). At bending frequencies of 0.5 to 2.0 Hz, bone formation rates increase as much as four-fold, while no increase is observed at frequencies lower than 0.5 Hz. When rest periods are inserted, the loaded bone shows increased bone formation rates when compared to bone subjected to a single bout of mechanical loading (147). Improved bone structure and strength is greatest if loading is applied in shorter versus longer increments (148). Therefore, for optimal

- 29- Chapter 2: Background anabolic loading, frequency, intensity, and timing of loading are all important parameters. The major challenge has been to translate these known in vivo parameters of mechanical loading to in vitro cell culture models.

2.2.2. Mechanisms whereby osteocytes sense mechanical loading

Even though osteocytes are thought to be mechanosensors, there is little conclusive data to show how mechanical loading is sensed by these cells. One of the more accepted forms of strain is the flow of bone interstitial fluid driven by extravascular pressure in combination with applied mechanical loading (149,150). Recently, the first real-time attempts to measure solute transport in bone through dye diffusion within the lacunar-canalicular system were conducted ex vivo (151). Fluid flow imposes a shear stress on osteocytes that appears to deform the cells within their lacunae and the dendrites within their canaliculi (150). Theoretical modeling predicts osteocyte wall shear stresses resulting from peak physiologic loads in-vivo in the range of 8 to 30 dynes/cm2. However, it is not clear if the dendritic processes, the osteocyte cell body, and/or cilia are the mechanosensors (Fig. 2.5).

Fig. 2.5: Cartoon showing potential ways that an osteocyte may sense fluid flow shear stress. (A). Fluid flow shear stress could perturb tethering elements between the canalicular wall and the cell membrane. (B). Fluid flow shear stress may also affect the cell body, causing cell deformation. (C). Fluid flow may perturb primary cilia leading to mechanosensation. Both matrix and cell deformation are also proposed to play a role in osteocyte mechanosensation. Bonewald LF 2006

A model of strain amplification in osteocyte cell processes has been proposed by Weinbaum and coworkers (152). One of the requirements of the model is that osteocyte dendritic processes be tethered to the canalicular wall and anchored to hexagonal actin bundles within the cell processes. The model predicts that fluid flow through this canalicular space will deform the shape of these

- 30- Chapter 2: Background tethering elements, creating a drag force that then imposes a hoop strain on the central actin bundles inside the osteocyte cell process. This model, however, does not take into account that the dendritic processes of osteocytes may not always be firmly anchored to their canaliculi. The osteocyte has been viewed as a quiescent cell until recently, when Dallas and co-workers showed cell body movement and the extension and retraction of dendritic processes (153). Calvarial explants from transgenic mice with green fluorescent protein (GFP) expression targeted to osteocytes were used to dynamically image living osteocytes within their lacunae. Surprisingly, these studies revealed that, far from being a static cell, the osteocyte may be highly dynamic. These data suggest that dendrites, rather than being permanent connections between osteocytes and with bone surface cells, may have the capacity to connect and disconnect. These studies also partially explain why a protein thought to play a role in dendrite elongation, E11/gp38, would be regulated by mechanical load in cells embedded in mineralized matrix (132). Fluid flow shear stress may induce mechanosensation in osteocytes through perturbation of integrins (154). Integrins, comprised of heterodimers of α and β subunits, are major receptors/ transducers that connect the cytoskeleton to the extracellular matrix (155) and interact with plasma membrane proteins such as metalloproteases, receptors, transporters, and channels mainly through the extracellular domain of their α subunits (156). The integrin α5 subunit may act as a tethering protein that, when perturbed by shear stress, opens hemichannels in osteocytes, allowing the release of prostaglandin (157). It has also been proposed that mechanical information is relayed in part by matrix and cell deformation (158–160). Typical in vivo strains in humans are on the order of 1,200 to 1,900 μE and were determined using strain gauges that covered an area approximately 1.8 mm by 3.6 mm; this area would contain thousands of cells and the strains measured are therefore averages. Microstructural strains measured at or near osteocyte lacunae were up to 3 times greater than the average strains measured with an external strain gauge (159,160). This suggests that the osteocyte is subjected to larger strains than those measured on the external bone surface. Recently, it has been shown that polycystin-1 and 2 (PKD1 and PKD2), known mechanosensory proteins in the kidney, do play a role in normal bone structure and that cilia do exist on both osteoblasts and osteocytes (161). Primary cilia clearly function as sensors of odors, light, and movement, depending on cell type (162). It remains to be determined whether the bone defect in animals with reduced or defective PKD1 function is due to defective mechanosensory function in bone cell cilia, as has been shown in kidney epithelial cells. Recently, Jacobs and coworkers provided preliminary data that loss of cilia resulted in decreased sensitivity to flow (163). It will be

- 31- Chapter 2: Background important to determine how a single cilium on an osteocyte cell body can mediate the mechanosensory functions ascribed to the osteocyte. In vivo, it has been shown that physiological loading prevents osteocyte apoptosis (164) and, conversely, that reduced mechanical loading in the tail suspension model increases osteocyte apoptosis (165). In vitro experiments have shown that fluid flow shear stress inhibits osteocyte apoptosis induced by serum starvation (166) and that substrate stretching prevents dexamethasone- induced apoptosis (167). Fluid flow shear stress has recently been shown to prevent both dexamethasone- and tumor necrosis factor-α-induced apoptosis, and this effect was shown to be mediated by prostaglandin production (168). Mechanical loading is therefore protective against apoptosis and this effect is mediated through prostaglandin production. From this reason, prostaglandin can now be added to the list of potential anti-apoptotic factors for osteocytes.

2.2.3. Osteocytes as regulators of mineralization and mineral metabolism

The osteoid-osteocyte may control deposition of mineral that begins to surround and encase this cell while it is embedding (169,170). It is also likely that this cell is subjected and responsive to loading. Mechanosensation may play a role in the process of selection of targeted osteoblasts on the bone surface to become osteocytes. Osteocytes in cortical bone are orderly and linearly arrayed. Signals passing from embedded cells to selected cells on the bone surface may be delivered through gap junctions to select a cell that will maintain this ordered network. Mature osteocytes also have the capacity to modify their local microenvironment. Glucocorticoid treatment causes mature osteocytes to enlarge their lacunae and remove mineral from their microenvironment (125). Osteocytes may be able to modify their microenvironment in response to other factors. Osteocytes may also play a major role in mineral homeostasis. Genes that are highly expressed in osteocytes are known regulators of mineralization and mineral homeostasis. The most convincing evidence that osteocytes are regulators of mineralization comes from studies of SOST/sclerostin. The SOST gene encodes a protein, sclerostin that is highly expressed in mature (not early) osteocytes and functions as an inhibitor of bone formation (171). The human conditions of sclerostosis and van Buchem disease are due to mutations in the SOST gene, and transgenic mice lacking sclerostin have increased bone mass. It appears that sclerostin is an indirect inhibitor of BMP, but specifically antagonizes the Wnt pathway (172) as an antagonist of LRP5, a gene shown to be important as a positive regulator of bone mass (173). Both Wnt/β-catenin and SOST are

- 32- Chapter 2: Background regulated by mechanical strain in osteocytes, positively and negatively, respectively. Is this one means by which loading regulates the bone formation and resorption responses? Deletion or mutation of genes that are highly expressed in embedding osteocytes and mature osteocytes, such as dentin matrix protein 1 (DMP1) and phosphate-regulating gene with homologies to endopeptidases on the X chromosome (PHEX), results in hypophosphatemic rickets (144,174). PHEX is a cell surface membrane metalloendoproteinase and DMP1 is expressed along the canaliculi of osteocytes. Other players in mineral metabolism include MEPE and FGF23, also highly expressed in osteocytes (175,176). Therefore, it has been proposed that the osteocyte network be viewed as an endocrine gland that can regulate mineral metabolism. DMP1, a promoter of mineralization and mineral homeostasis, and MEPE, an inhibitor of mineralization, both increase sequentially in response to mechanical load (130). This raises the question whether mineral metabolism could be regulated by mechanical loading. Another level of complexity, but a challenge for further investigation.

2.3. Bone Remodeling

Bone remodeling, a complex process by which old bone is continuously replaced by new tissue, requires interaction between different cell phenotypes and is regulated by a variety of biochemical and mechanical factors allowing the maintenance of the shape, quality, and size of the skeleton (3). This is accomplished through the repairing of microfractures and the modification of structure in response to stress and other biomechanical forces. This process is characterized by the coordinated actions of osteoclasts and osteoblasts, organized in bone multicellular units (BMU) that follow an activation-resorption-formation sequence of events. In a homeostatic equilibrium, resorption and formation are balanced so that old bone is continuously replaced by new tissue and adapts to mechanical load and strain (40). In cortical bone the BMU forms a cylindrical canal about 2,000 µm long and 150–200 µm wide and gradually burrows through the bone with a speed of 20–40 µm/day. During a remodeling cycle, osteoclasts dig a circular tunnel in the dominant loading direction (41) and then are followed by several thousands of osteoblasts that fill the tunnel (42). In this manner, between 2% and 5% of cortical bone is being remodeled each year. The trabecular bone is more actively remodeled than cortical bone due to the much larger surface to volume ratio. Osteoclasts travel across the trabecular surface with a speed of approximately 25 µm/day, digging a trench with a depth of 40–60 µm. In a remodeling cycle resorption begins with the migration of partially differentiated mononuclear preosteoclasts to the bone surface where they form multinucleated osteoclasts. After the completion

- 33- Chapter 2: Background of osteoclastic resorption, there is a reversal phase when mononuclear cells appear on the bone surface. These cells presumably prepare the surface for new osteoblasts to begin bone formation and provide signals for osteoblast differentiation and migration. The formation phase follows with osteoblasts laying down bone until the resorbed bone is completely replaced by new mineralized matrix. When this phase is complete, the surface is covered with flattened lining cells and a prolonged resting period begins until a new remodeling cycle is initiated. The stages of the remodeling cycle have different lengths. In humans, resorption continues for about 2 weeks, the reversal phase may last up to 4 or 5 weeks, while formation can continue for 4 months until the new bone structural unit is completely created (46).

2.3.1. Regulation of Bone Remodeling

The overall integrity of bone appears to be controlled by hormones and many other proteins secreted by both hemopoietic bone marrow cells and bone cells. There is both systemic and local regulation of bone cell function:

1) Systemic regulation – gonadal hormones (estrogens and androgens) are the most important regulators of bone remodeling. The majority of postmenopausal women show a marked decrease in bone mineral density and high-turnover bone metabolism. This phenomenon leads to postmenopausal osteoporosis (184,185). Experimentally-induced estrogen deficiency by ovariectomy in female animals causes similar bone alterations. When estrogen-deficient animals and postmenopausal women are treated with exogenous estrogen, the decrease in bone mass and increase in bone turnover are reversed. Estrogens decrease the responsiveness of the osteoclast progenitor cells to RANKL, thereby preventing osteoclast formation (53). Furthermore, besides reducing osteoclast life span (54) estrogens stimulate osteoblast proliferation and decrease their apoptosis. They affect gene coding for enzymes, bone matrix proteins, hormone receptors, transcription factors, and they also up-regulate the local production of OPG, IGF-I, IGF-II, and TGF-β (55). This suggests that estrogens have a bone-protective effect. Estrogen effects are mediated via estrogen receptor α and/or β (ERα, ERβ) and receptor function is species and gender specific (186). ERα mainly functions in various estrogen target organs (187), however, ERα knockout (ERαKO) mice exhibit increased bone volume: tissue volume ratios (BV/TV) regardless of gender (188). ERαKO mice have low-turnover bone metabolism. The numbers of osteoclasts and osteoblasts are reduced, and both the bone resorption and bone formation rates are slower, as determined by a bone morphometric analysis. Trabecular bone mineral density (BMD) is increased in male but not in female ERαKO mice. Testosterone levels are markedly increased regardless of

- 34- Chapter 2: Background sex. In contrast, ERβKO mice have a higher BV/TV, though BMD is similar to that in wild-type mice. In brief, female ERKO mice including ERα and ERβ double knockout mice do not exhibit bone loss characteristic of postmenopausal osteoporosis in humans (189). That estrogen has bone protective effects in the human is virtually indisputable. Despite studies of bone metabolism in ERKO mice, however, the mechanism behind these effects has remained elusive up until the generation of osteoclast-specific ERαKO mice as described below. Estrogen deficiency is also important in men. The increase in BMD in young men and its decrease with aging is related to circulating free estrogen, rather than testosterone. There is evidence to suggest that estrogens regulate bone resorption and that both estrogen and testosterone regulate bone formation in men (185). On the other hand, male bone has a higher mineral density and lower risk for fracture or osteoporosis compared to that of the female (190). The greater strength of male bone has been attributed to the anabolic effect of androgenic hormones. Androgens are essential for skeletal growth and maintenance via their effect on androgen receptor, which is present in all types of bone cells (56). Androgens are synthesized from cholesterol through several enzymatic pathways in which the side chain of cholesterol is shortened through oxidation from 27 carbons to 19 carbons (191). In men, androgens are secreted almost exclusively from the testes as testosterone. The adrenal glands also secrete dehydroepiandrosterone (DHEA), which is a minor androgen that also serves as a substrate for peripheral aromatization to estradiol (E2). Testosterone is either converted by 5a-reductase to dihydrotestosterone (DHT), or metabolized to E2 by aromatase, a widely distributed microsomal cytochrome P450 enzyme. The former pathway amplifies androgen action locally while the latter pathway diversifies androgen action (191). Hence, enzymatic androgen activation leads to testosterone acting directly or via its more potent metabolite DHT through the androgen receptor (AR), or indirectly via aromatization to E2 through the estrogen receptors (ERs). Thus, testosterone functions as a precursor for peripheral conversion into biologically highly active hormones. Estradiol, which is thought to play a major role in bone metabolism in men, is largely synthesized by extratesticular aromatization of circulating testosterone with only a small proportion of E2 (approximately 15–20%) being directly secreted by the testes (192). Depending on the relative activity of aromatase, 5a-reductase, and dehydrogenases, and the relative distribution of ARs and ERs in peripheral target tissues, testosterone and its metabolites may predominantly activate either the AR or the ER. In bone tissue, the expression of aromatase (193), 5α-reductase (194), 17beta-hydroxysteroid dehydrogenase (17β-HSD (195)), and 3β-HSD (196) has been documented, supporting the concept of tissue-specific peripheral activation of gonadal hormones.

- 35- Chapter 2: Background The AR has been identified inmost bone cells, including osteoblasts (197), osteocytes (198), and osteoclasts (199). Estrogen action on bone, in men and women, is mediated via ERs. These nuclear hormone receptors are also expressed in osteoblasts, osteoclasts, and osteocytes (200). Two ERs have been identified: ERa is predominantly expressed in cells resident in cortical bone, whereas ERb shows higher levels of expression in cells found in cancellous bone (201). Alternate, non- genomic pathways have also been described in which ARs and ERs modulate transcription indirectly, via protein–protein interactions. Indeed, AR knockout (ARKO) mice exhibit high bone turnover with increased bone resorption, which results in reduced trabecular and cortical bone mass without affecting bone morphology. Bone loss in gonadectomized male ARKO mice is only partially prevented by treatment with aromatizable testosterone. Examination of primary cultured osteoblasts and osteoclasts of ARKO mice reveals that AR function is necessary for the suppressive effects of androgens on osteoclastogenesis (185). Whether the bone-forming osteoblast or bone- resorptive osteoclast is the direct target of androgen-AR signaling remains to be clarified. Parathyroid hormone (PTH) is also thought to play an important role in calcium homeostasis. It maintains serum calcium concentrations by stimulating bone resorption, increasing renal tubula calcium reabsorption and renal calcitriol production. PTH stimulates bone formation when given intermittently and bone resorption when secreted continuously (48). Calcitriol is essential in enhancing intestinal calcium and phosphorus absorption, and in this way it promotes bone mineralization. In addition, vitamin D3 possesses important anabolic effects on bone, thus exerting a dual effect on bone turnover (49). Calcitonin, in pharmacologic doses, mediates loss of the ruffled border, cessation of osteoclast motility, and inhibition of the secretion of proteolytic enzymes through its receptor on osteoclasts. This effect, however, is dose limited and its physiologic role is minimal in the adult skeleton. The growth hormone (GH)/IGF-1 system and IGF-2 are important for skeletal growth, especially at the cartilaginous end plates and during endochondreal bone formation. They are among the major determinants of adult bone mass through their effect on regulation of both bone formation and resorption (50). Glucocorticoids exert both stimulatory and inhibitory effects on bone cells. They are essential for osteoblast maturation by promoting their differentiation from mesenchymal progenitors but they decrease osteoblast activity. Furthermore, glucocorticoids sensitize bone cells to regulators of bone remodeling and they augment osteoclast recruitment (51). Thyroid hormones stimulate both bone resorption and formation. Thus, bone turnover is increased in hyperthyroidism and therefore bone loss can occur (52).

- 36- Chapter 2: Background The discovery that the brain controls bone remodelling has provided a new paradigm for understanding of neuroskeletal biology. Recently, the discovery of central control of bone mass by leptin shed light on a novel pathway controlling bone mass: osteoporosis is considered to be not just a bone disease, but also a ‘neuronal’ disease (202). Successively, other neuropeptides such as neuropeptideY (NPY), cocaine- and amphetamine-regulated transcript (CART) and, more recently, neuromedin U (NMU) have been demonstrated to be bone-regulating neuropeptides (203–205). Leptin is a 16-kDa peptide hormone synthesised by adipocytes that affects appetite and energy metabolism through its binding to the leptin receptor located in the hypothalamus (206). ob⁄ob mice that lack functional leptin are obese and sterile (6). In spite of hypogonadism, the most common cause of osteoporosis, ob ⁄ ob mice and db ⁄db mice that lack a functional leptin receptor display high bone mass (207). Importantly, ob⁄ob mice fed a low fat diet have a normal weight and high bone mass (207). Moreover, mouse models of lypodystrophy, such as A-ZIP transgenic mice expressing a dominantnegative protein inhibiting B-ZIP adipocyte transcription factors or Pparchyp ⁄ hyp mice carrying a hypomorphic mutation at the PPARc2 locus (208), have decreased leptin serum levels, due to low fat, and display high bone mass. Thus, regardless of their body weight, low serum leptin level induces an increase of bone mass, demonstrating that a leptin-signalling defect is the bona fide cause of high bone mass. ob/ob mice have a higher bone formation rate with a concomitant increase in bone resorption (207). There are two regions in the hypothalamus, namely arcuate (Arc) nuclei and ventromedial hypothalamic (VMH) nuclei, which are rich in leptin receptors (209). Destruction of Arc by monosodium glutamate in wild-type mice induces obesity, but not high bone mass. By contrast, destruction of VMH by gold thioglucose in wild-type mice recapitulates the bone phenotype of ob⁄ob mice; high bone mass due to an increase in bone formation. More importantly, i.c.v. leptin infusion to VMH-lesioned ob ⁄ ob mice decreases body weight, but does not affect bone mass. Conversely, i.c.v. leptin infusion to Arc-lesioned ob⁄ob mice decrease bone mass, but does not affect body weight. These results suggest that VMH neurones are necessary for the leptin-dependent central regulation of bone mass (210). Along with its anorexigenic effect, leptin exerts various physiological roles including sympathetic nervous system (SNS) regulation (211). For example, the sympathetic tone of ob ⁄ ob mice is low and i.c.v. leptin infusion increases catecholamine secretion (212). In addition, the stereotactic infusion of leptin to VMH nuclei, and not other nuclei, induces SNS activation (213). Along with these observations, many osteoblasts reside next to sympathetic neurones in bone marrow and also express the beta2- adrenergic receptors (adrb2) specifically, indicating the interaction of SNS and bone remodelling (210). Indeed, mice treated with isoproterenol, a betaagonist, display a massive decrease in bone

- 37- Chapter 2: Background mass and mice that are blocked SNS signalling, either genetically (adrb2-deficient mice or dopamine b-hydroxylase-deficient mice) or pharmacologically (wildtype mice treated with a beta- blocker), all exhibit a high bone mass phenotype due to an increase in bone formation (204,210). These mice are also protected from the inhibition of bone formation by leptin. Thus, SNS is a major, if not the only, pathway that is responsible for the inhibitory role on bone formation by leptin. Leptin and SNS also regulate osteoclastic resorption (204). In addition to increased bone formation, adrb2-deficient mice also display decreased bone resorption and possess fewer numbers of osteoclasts. Serotonin is an indoleamine produced in enterochromaffin cells of the duodenum and in serotonergic neurons of brainstem that does not cross the blood brain barrier (214). Thus, it is de facto a molecule with two distinct functional identities depending on its site of synthesis: a hormone when made in the gut and a neurotransmitter when made in the brain (215,216). Study by Yadav et al. (216) showed that brain-derived serotonin (BDS) promotes bone mass accrual when acting as a neurotransmitter. The central function of serotonin is mediated through the Htr2c receptor expressed in ventromedial hypothalamic neurons (VMH). Htr2c_/_ mice are markedly osteopenic before any metabolic modification is detectable, indicating that serotonin regulation of bone mass occurs independently of its effects, through Htr2c, on energy metabolism. Neuropeptide Y receptor (NPY) is expressed in the central and peripheral nervous system and has been shown to exhibit various physiological actions including food intake regulation. To date, five receptors (Y1, 2, 4, 5 and 6) have been identified as NPY receptors (217). Of these, Y1 and Y5 are considered important for appetite regulation through the analysis of knockout mice (217). Y2- deficient or hypothalamic specific Y2-deficient mice develop a high bone mass phenotype accompanied by an increase in bone formation, demonstrating that Y2 signalling affects bone formation through the CNS (218). Recently, Y1-deficient mice have also been shown to display high bone mass due to an increase in bone formation (203). However, hypothalamus-specific Y1 deletion does not affect bone mass, suggesting that the nature of Y1 receptor signalling affecting bone remodelling is peripheral. Interestingly, germline deletion of Y2 significantly reduced the expression of Y1 in osteoblasts, indicating that high bone mass in Y2-deficient mice may be attributable to that, at least in part (219). Although Y4-deficient mice have normal bones, Y2⁄Y4 double mutant mice display a higher bone mass and lower serum leptin level than Y2 single mutant mice, suggesting an indirect effect of bone remodelling through leptin-signalling by Y4 (220). To date, there is no evidence of the interaction between Y receptor signalling and SNS for bone remodelling.

- 38- Chapter 2: Background The cannabinoid system, known to regulate analgesia, appetite and energy expenditure, has also been shown to regulate bone mass in vivo. There are two cannabinoid receptors: CB1, encoded by the CNR1 gene, is predominantly expressed in the CNS and SNS, as well as peripheral tissues, and CB1 is responsible for most of the actions of the CNS with respect to cannabinoid drugs and endocannabinoids. By contrast, CB2 is more specific for peripheral tissues, including osteoblasts and osteoclasts. CNR1-deficient mice on an outbred CD1 background exhibit high bone mass with normal bone formation and resorption (221), whereas CNR1-deficient mice on an inbred C57Bl ⁄6J background display a low bone mass associated with a decrease in bone formation and an increase in osteoclast number (222). The molecular basis for this discrepant phenotype is unknown, but it is interesting that CNR1-deficient mice on a C57Bl ⁄6J background are hypersensitive to i.c.v. leptin, which explains the low bone mass phenotype, at least in part. By contrast, CNR2-deficient mice display a low bone mass phenotype with an increase in bone formation and in osteoclast number (223). The fact that CB2 agonists stimulate osteoclast formation in vitro indicates that these compounds act directly on osteoclasts (223). A growing number of studies describing ‘unexpected’ bone phenotypes in mutant mice deficient for neuropeptides or neurotransmitters have now established a new research area linking skeletal and neuronal biology.

2) Local regulation - as far as the local regulation of bone cell function is concerned, the recent discovery of the OPG/RANKL/RANK system, has given a clearer picture regarding the control of osteoclastogenesis and bone remodeling in general. RANKL, expressed on the surface of preosteoblastic/stromal cells and a subsets of T-cells (47) binds to RANK on the osteoclast precursor cells and is critical for the differentiation, fusion into multinucleated cells, activation, and survival of osteoclastic cells (43). OPG inhibits the entire system by competitively binding to RANKL (44,45). Macrophage colony-stimulating factor (M-CSF), which binds to its receptor, c- Fms, on preosteoclastic cells is also necessary for osteoclast development (57). The opposite phenotypes of OPG overexpression or with RANKL deletion mice (osteopetrosis) and OPG-deficient or with RANKL overexpression (osteoporosis), have led to the hypothesis that OPG and RANKL can be the mediators for the stimulatory or inhibitory effects of a variety of systemic hormones, growth factors, and cytokines on osteoclastogenesis. This has been referred to as “the convergence hypothesis” in that the activity of the resorptive and antiresorptive agents “converges” at the level of these two mediators, whose final ratio controls the degree of osteoclast differentiation, activation, and apoptosis (58).

- 39- Chapter 2: Background A number of cytokines such as TNF-α and IL-1 modulate this system primarily by stimulating M- CSF production and by directly increasing RANKL expression (59). In addition, a number of other cytokines and hormones exert their effects on osteoclastogenesis by regulating cell production of OPG and RANKL (60-66). Furthermore, IL-6, a pleiotropic cytokine secreted by osteoblasts, osteoclasts, and stromal cells, appears to be an important regulator of bone remodeling by stimulating osteoclastic bone resorption (67) but also by promoting osteoblast generation in conditions of high bone turnover (68). Recent studies have also suggested that osteoblast-derived PTHrP promotes the recruitment of osteogenic cells and prevents the apoptotic death of osteoblasts, thus being an important regulator of bone cell function (69). Abnormalities of bone remodeling can produce a variety of skeletal disorders, mainly osteoporosis the most abundant degenerative disease in western societies. The recent advances concerning systemic and local regulation of bone remodeling have led to new approaches in the diagnosis and treatment of these disorders. In particular, the newer methods in molecular and cellular biology aid the definition of the abnormalities in cells of the osteoblastic and osteoclastic lineages that lead to bone disease and the development of new therapeutic approaches based on a better understanding of the pathogenetic mechanisms. These involve production of recombinant molecules of cytokines and their soluble receptors, development of inhibitory peptides, and specific inhibition of key signaling pathways.

2.4. Mechanical load induced bone adaptation

The skeleton is able to continually adapt to mechanical loading by adding new bone to withstand increased amounts of loading, and by removing bone in response to unloading or disuse (reviwed in (126,127)). Galileo, in 1638, is documented as first suggesting that the shape of bones is related to loading. Julius Wolff, in 1892, more eloquently proposed that bone accommodates or responds to strain. Mechanical loading is perhaps the most important single physiological/environmental factor regulating bone mass and shape. Although the basic form and development of bone are genetically encoded, their final mass and architecture are governed by adaptive mechanisms sensitive to the mechanical environment. Mechanical signals are transmitted to bone mainly by muscle contractions generating strains in the bone matrix (70). Loss of bone (osteoporosis) and muscle strength (sarcopenia) develop together with increasing age (71); characteristic to osteoporosis is the failure of structural adaptation by bones to the mechanical environment, which results in increased

- 40- Chapter 2: Background incidence of fractures in response to physiological loads or minimal trauma. Although age-related bone loss cannot be ascribed entirely to sarcopenia (71), a growing number of studies in humans report that resistance training is an effective means of preserving and increasing the mass of both muscle and bone at all ages (72-74). Likewise, a handful of studies in experimental animals have demonstrated a mechanical load-induced stimulation of bone formation. Anatomically, the crucial structural component of all major skeletal load-bearing sites, namely proximal femur, vertebrae and distal radius, is trabecular bone. During growth increasingly vigorous mechanical usage increases global bone deposits by enhancing longitudinal growth through the addition of new spongiosa and new cortex, in addition to stimulating cortical modeling drifts of increased cortical cross-sectional area. However, in the adult organism, in which modeling drifts are usually ineffective and cortical bone turnover is relatively low, the effects of vigorous mechanical usage are targeted mainly to the spongiosa and endosteal cortical surfaces where losses and marrow cavity expansion are retarded (75). Furthermore, significant gains in trabecular bone mass have been reported in exercising healthy humans (76). By contrast, decreased mechanical usage results in increasing numbers of BMUs and high bone turnover, with a clear shift from a balance between bone resorption and formation towards increased resorption and decreased formation (77). These observations in humans have been repeatedly supported by experimental work in laboratory animals, thus confirming Frost’s mechanostat theory (75). This theory defines four mechanical usage windows, with thresholds defined by minimum effective strains (MES): (i) trivial (subphysiological) loads which result in a negative, high trabecular bone turnover; (ii) physiological loads responsible for normal, balanced turnover; (iii) overload, which induces positively balanced turnover; and (iv) pathological excessive loading, or failure loads, which result in microfractures and in addition to enhanced lamellar bone formation produce woven bone, apparently as part of the fracture healing process (75,78). The effect of trivial loads has been confirmed in models employing immobilization, by methods such as schiatic denervation (79), limb fixation (80), hypogravity (81- 83) and tail suspension (84,85). Decreased loading also occurs in joint injuries (86), and the effect on bone in the joint region results in decreased bone volume (87,88), primarily through architectural adaptation (89). The effect of overloading has been studied using a wide variety of approaches. For example, increased bone formation indices have been reported in animals forced into excessive exercising regimens (90). A 6-fold increase has been reported in woven spongiosa formed in vitro in intermittently pressurized hydraulic bone chambers (91). Woven bone is also produced on trabeculae in response to extraordinary loading conditions (92-94). Lower load magnitudes produce increases in trabecular lamellar bone (94-97). Mechanical loading also reduced ovariectomy

- 41- Chapter 2: Background induced loss of metaphyseal spongiosa (80); at least in rats, the respective signaling and anabolic effects generated by overloading are parathyroid hormone and estrogen dependent (98-101). Furthermore, trabecular anabolic bone adaptation can be affected by specific load profiles such as high frequency, low magnitude vibration (97), and is reflected both in bone architecture (102) and mechanical strength (103). Judex et al. (224) have demonstrated that extremely small magnitude forces, induced non-invasively to the skeleton as whole body vibrations, can be perceived as osteogenic (233). In the proximal tibia of adult BALB/cByJ mice, for example, 10 min per day of a 45 Hz, 0.3g acceleration (1g = acceleration on Earth, or 9.8 m/s2) was anabolic to trabecular bone, while disuse was catabolic and may also suppress bone formation (238). This study reported relation of these mechanically mediated changes in bone formation rates (BFR) to the expression of a broad set of genes anticipated to play a role in regulating bone adaptation. The thirteen genes considered all have critical, but not necessarily unique, tasks in (mechanically induced) bone formation (Cbfa1 (225), osterix (226), BMP-2 (227), IGF-1 (228), MMP-2 (229), collagen type I (230), integrin b3 (231), osteonectin 232)) and bone resorption (RANKL(234), iNOS (235), osteopontin (236), MMP-9 (237), cathepsin K (237)). They hypothesized that alterations in load bearing (increase or decrease) will stimulate differential responses in the activity of these formation and resorption gene ‘‘families,’’ including their temporal expression patterns. These data emphasized that the molecular events involved in mechanically mediated bone adaptation was both subtle and complex. Further, the similarity in expression patterns between many distinct genes responding to the catabolic and/or anabolic signals accentuated an intricate co-dependence of molecular events involved in bone’s adaptation to mechanical signals. Another studies by Roling et al. (239) showed how long bones was adapted in response to loading. Cyclic mechanical loads were applied axially along the ulna of adult rats three times per week for 16 weeks. The rat ulna has a natural curvature in the medial-lateral direction, so axial loads induced bending of the bone. Under load, the medial surface of the bone was subjected to compressive stresses and the lateral surface was in tension. The ratio of compressive to tensile stress magnitude in the loaded adult rat ulna was about 1.5, indicating that the highest stress in the ulna occured at the medial surface in compression. The pattern of bone formation induced by loading resembles the stress distribution, with more bone formation where the stresses were highest. The improvement in bone structure was evidenced by a 69% increase in second moment of area. The ulnar bone strength of loaded limbs was 64% greater than controls and energy absorbed before fracture increased by 94%, yet the improvement in bone mineral content (BMC) was only a modest 7%. In this animal

- 42- Chapter 2: Background model, loading induced dramatic improvements in bone biomechanical properties, even with small changes in BMC. The structural efficiency of the ulna was improved by bone formation, specifically in highly stressed areas where it was most needed. A variety of in vitro models have been proposed to study the cellular and molecular mechanisms involved in the anabolic effect of loading. However, the relevance of these models to the in vivo situation is equivocal mainly due to the inability for definitive identification of mechano-sensitive cells (e.g., osteocytes, osteoblasts, lining cells) and the absence of pathways by which loads applied to the cortical envelope are transferred to trabecular bone cells (e.g., cell and cell attachment molecule deformation, fluid flow (106)). Still, a few studies have assigned a mechanosensitive role to osteocytic, periosteal and endosteal cells by demonstrating changes in signalling molecules such as integrins and the glutamate receptor, tanacin-C, and RoBo-1 (107-110). However, these studies were carried out in cortical bone only and their relevance to trabecular bone cells remains to be investigated. It is known that the effect of stresses applied at different rates at an object is largely determined by the material properties of that object. Low magnitude (<10με) and high frequency (10–100 Hz) loading can stimulate bone growth and inhibit disuse osteoporosis, while high loading rates have been shown to increase bone mass and strength after jumping exercises in middle-age osteopenic ovariectomized rats (249). For bone cells, Bacabac and colleagues [29–31] have shown that the production of signaling molecules in response to an in vitro fluid shear stress (at 5 and 9 Hz) and vibration stress (5–100 Hz) correlated with the applied stress rate (241–243). The faster the stress was applied, the stronger the observed response of the cells (244), suggesting that the bone cellular response to loading and mechanical properties of the cell are related, which implies that the response of bone cells to loading is related to cytoskeletal properties. The same group developed a novel application of two- particle microrheology, for which a 3D in vitro system was devised to quantify the forces induced by cells on attached fibronectin-coated probes (4μm). The frequency at which the cells generate forces on the beads is related to the metabolic activity of the cell (245). With this device and using NO production as a read-out, the material properties of round suspended MLO-Y4 osteocytes and flat adherent MLO-Y4 osteocytes were characterized. Osteocytes with round suspended morphology required lower force stimulation in order to show an increase in NO production, even though they were an order-of-magnitude more elastic compared to flat adherent cells (246). Apparently, elastic osteocytes seem to require less mechanical forces in order to respond than stiffer cells (246). In contrast, flat adherent MLO-Y4 osteocytes, primary chicken osteocytes, MC3T3-E1 osteoblasts, and primary chicken osteoblasts all showed a similar elastic modulus of less than 1 kPa

- 43- Chapter 2: Background (245). This indicates that differences in mechanosensitivity between osteocytes and osteoblasts might not only be directly related to the elasticity of the cell, but also to other cell-specific properties, i.e., the presence of receptors or ion channels in the membrane, or how cells change their material properties in relation to deformation. Studies addressing load-induced cellular and molecular mechanisms in trabecular bone are rather few. Compared to cortical bone, the spongiosa is enclosed in the cortical envelope and is substantially less accessible to cell isolation techniques, particularly in small laboratory animals such as rats and mice. Similar considerations also apply to the assessment of mechanical load induced trabecular deformation. Of particular relevance to the present thesis is the structural modeling of trabecular deformation generated by force applied to the vertebral cortex using micro- tomographic imaging (μCT) (111) and work emanating from the Chambers group. The latter investigators have devised a rat model in which the body of the eighth caudal vertebra is subjected to controlled, atraumatic cyclic compression administered via pins introduced into the bodies of the seventh and ninth vertebrae (99,112). Using electron microscopy and in situ hybridization they have recently demonstrated activation of trabecular lining cells and an early (1 h) increase in c-fos and IGF-I expression in trabecular osteocytes. Apparently, the c-fos response is associated with Ca2+ signaling pathways and integrin binding (113). A late (72 h) increase in transcripts for collagen type I and osteocalcin was observed in trabecular osteoblasts/lining cells (114-116). While contributing important information, by and large these data remain incomplete inasmuch as it is very unlikely that increases in c-fos and IGF-I transcripts comprise the entire inter- and intracellular signaling cascade evoked by overloading. Nevertheless, these and a few other studies (108,116-118) assigned for the first time an experimentally supported physiologic role for osteocytes and lining cells, which the present project proposes to further substantiate and define in mice (119). Recently, Webster et al. (120,183) devised a mouse tail loading device (CVAD) and investigated the trabecular bone adaptation by cyclic mechanical stimulation of the fifth caudal vertebrae (C5) of C57BL/6 (B6) female mice. They reported a genotype dependent dose response in trabecular bone in several microarchitectural parameters following regular bouts of mechanical stimulation. Mice were randomly divided into 4 loading groups: 0N (sham-loaded), 2N, 4N and 8N. Using the CVAD the C5 of all mice were subject to an acute loading regime (3000 cycles, 10 Hz, 3 times a week for 4 weeks). Analysis of microCT image and histomorphometry data revealed that trabecular bone formation was successfully induced. At trabecular sites global bone volume density (BV/TV) increased by 25.9% in B68N loading group comparing to sham-loaded (P < 0.05). This was

- 44- Chapter 2: Background accompanied by a significant global increase of 21.9% (P < 0.001) in trabecular thickness (Tb.Th) and local significant increases of up to 14.6% (P < 0.001) in trabecular number (Tb.N). The findings show that trabecular bone in both biological strains is responsive to mechanical loading and that in terms of absolute and percentage increases, B6 mice are more mechano-sensitive for a given load. This study has established a mouse model which, for the first time, will allow the study of load regulated gene expression in both trabecular and cortical bone; furthermore it has also demonstrated the potential candidacy of B6 and C3H mice for a functional genomics approach at isolating the mechano-sensitive gene(s) specific to both cortical and trabecular bone.

2.5. Mouse Genetics

With so many aspects contributing to the strength of bone discovering the genes responsible presents a sizeable task and cannot be done by studying human biology alone. The study of genetics in humans is limited to some degree by the tremendous heterogeneity among population, as well as multiple genetic, heritable and environmental determinants of the target phenotype. If the challenge is to be realized a more controllable genetic model is required, which is why the mouse has become such an important tool. The biological similarities between man and mouse make this small animal the ideal experimental surrogate and via the field of comparative genomics this small animal greatly increases the likelihood that candidate genes and effective therapies will be found (178). This section provides an overview about current efforts in mouse genetics. It is adapted from Silver et al. (179). Many features of human biology at the cell and molecular levels are shared across the spectrum of life on earth; our more advanced organism-based characteristics are shared in a more limited fashion with other species. At one extreme are a small number of human characteristics (brain functions and behavior) that are shared by no other species or only by primates. But at a step below there is a whole set of characteristics, which are shared only with mammals. In this context, the importance of mice in genetic studies was first recognized in the biomedical fields of immunology and cancer research, for which a mammalian model was essential. Although it has been obvious that many other aspects of human biology and development should be amenable to mouse models, until recently, the tools just did not exist to allow for a genetic dissection of these systems. The movement of mouse genetics to the forefront of modern biomedical research was catalyzed by the recombinant DNA revolution, which began 30 years ago. With the ability to isolate cloned copies of genes and to compare DNA sequences from different organisms came the realization that

- 45- Chapter 2: Background mice and humans as well as all other placental mammals are even more similar genetically than they were thought to be previously. An astounding finding has been that all human genes have counterparts in the mouse genome which can almost always be recognized by cross-species hybridization. Thus, the cloning a of human gene leads directly to the cloning of a mouse homolog which can be used for genetic, molecular, and biochemical studies that can then be extrapolated back to an understanding of the function of the human gene. Although the haploid chromosome number associated with different mammalian species varies tremendously, the haploid content of mammalian DNA remains constant at approximately three billion base pairs. It is not only the size of the genome that has remained constant among mammals; the underlying genomic organization has also remained the same as well. Large genomic segments (on average, 10-20 million base pairs) have been conserved virtually intact between mice, humans, and other mammals. In fact, the available data suggest that a rough replica of the human genome could be built by simply breaking the mouse genome into 130-170 pieces and pasting them back together again in a new order (180,181). Although all mammals are remarkably similar in their overall body plan, there are some differences in the details of both development and metabolism, and occasionally these differences can prevent the extrapolation of mouse data to humans and vice versa (182). Nevertheless, the mouse has proven itself over and over again as being the model experimental animal par excellence for studies of nearly all aspects of human genetics. Besides the strong homology in the genome, among mammals the mouse is ideally suited for genetic analysis for several other reasons too. First it is one of the smallest mammals known, second it has a short generation time, in the order of 10 weeks from being born to giving birth. Third, females breed prolifically in the laboratory with an average of 5-10 pups per litter. Fourth, an often forgotten advantage is the fact that fathers do not harm their young and that laboratory-bred strains are relatively docile and easy to handle. Finally, investigators are even able to control the time of pregnancies.

Manipulation of the mouse genome and micro-analysis The close correspondence discovered between the genomes of mice and humans would not have been sufficient to drive researchers into mouse genetics without the simultaneous development, during the last decade, of increasingly more sophisticated tools to study and manipulate the embryonic genome. Today, genetic material from any source (natural, synthetic or a combination of the two) can be injected directly into the nuclei of fertilized eggs; two or more cleavage-stage embryos can be teased apart into component cells and put back together again in new "chimeric"

- 46- Chapter 2: Background combinations; nuclei can be switched back and forth among different embryonic cytoplasma; embryonic cells can be placed into tissue culture, where targeted manipulation of individual genes can be accomplished before these cells are returned to the embryo proper. Genetically altered live animals can be obtained subsequent to all of these procedures, and these animals can transmit their altered genetic material to their offspring. Progress has also been made at the level of molecular analysis within the developing embryo. With the polymerase chain reaction (PCR) protocol, DNA and RNA sequences from single cells can be characterized and enhanced versions of the somewhat older techniques of in situ hybridization and immunostaining allow investigators to follow the patterns of individual gene expression through the four dimensions of space and time. Finally, with the automation and simplification of molecular assays that has occurred over the last several years, it has become possible to determine chromosomal map positions to a very high degree of resolution. Genetic studies of this type are relying increasingly on extremely polymorphic microsatellite loci to produce anchored linkage maps, and large insert cloning vectors, to move from the observation of a phenotype to a map of the loci that cause the phenotype, to clones of the loci themselves. All of these techniques provide the scientific community with the ability to search for answers to the many questions posed. This will invariably lead to more questions, but the potential is there to elucidate the mechanisms of many diseases and realize effective treatments.

The mouse and osteoporosis

Rodent models for testing hypotheses to skeletal disorders are not new. In fact this is how many of the established treatments came to market. The overiectomised rat is a well established tool and was used to test how estrogen deprivation affects the bone remodeling unit. At the forefront of technology today is the mouse model. Numerous mouse models exist, each of which attempt either to identify or evaluate candidate genes associated with osteoporosis.

Differential gene expression arrays

The use of cDNA microarrays offers substantial advantage for the study of load-induced molecular changes and signaling pathways in bone. A DNA microarray is a multiplex technology used in molecular biology and in medicine.The cDNA microarrays have revolutionized the way in which gene expression is now analyzed, by allowing the RNA product of thousands of genes to be monitored at once. By examining the expression of so many genes simultaneously it is now possible to identify and study gene expression patterns that underlie cellular physiology. For example

- 47- Chapter 2: Background scientists can now see which genes are switched on (or off) as cells grow, divide or respond to hormones or toxins. DNA microarrays are little more than glass microscope slides studded with a large number of DNA fragments, each containing a nucleotide sequence that serves as a probe for a specific gene. The mostly dense arrays may contain tens of thousands of these fragments in an area smaller than a postage stamp. These arrays are generated from DNA probes which have been produced by RT-PCR and then spotted onto the slide by a robot thus the exact sequence and position of every probe on the array is known. Any nucleotide fragment that hybridizes to a probe on the array can now be identified as the product of a specific gene simply by detecting the position to which it is bound.

DNA probes describing RNA from sample 2, labeled known genes are amplified with green fluorochrome using PCR and printed into a matrix on a glass slide

Hybridization

RNA from sample 1, labeled with red fluorochrome

Fig. 2.8: Use of DNA microarrays to monitor the expression of thousands of genes simultaneously (Alberts et al. Molecular biology of the cell. 4 ed. 2002).

To use DNA microarrays to monitor gene expression RNA from the cells being studied is extracted and converted to cDNA. The cDNA is then labeled with a fluorescent probe. The microarray is incubated with this labeled cDNA sample and hybridization is allowed to occur (Fig. 2.8). The array is then washed to remove cDNA that is not tightly bound, and the positions in the microarray to which labeled cDNA fragments have bound are identified by an automated scanning-laser microscope. The array positions are then matched to the particular gene whose sample DNA was spotted in this location. In figure 2.9 RNA has been collected from two different cell samples for a direct comparison of their relative levels of gene expression. Theses samples are labeled, one with a red fluorochrome, and the other with a green fluorochrome. Hence red spots in the hybridized array indicate that the gene in sample 1 is expressed at a higher level than the corresponding gene in sample 2. Green spots indicate that expression of the gene is higher in sample 2 than in 1. Yellow spots indicate that the genes are expressed in equal amounts in both samples while the dark spots indicate little or no expression in either sample.

- 48- Chapter 2: Background References 1. Alberts, B., Johnson A, Lewis J, Raff M, Roberts K, Walter P 2002 Molecular Biology of the Cell, 4 ed. Garland Science, New York.

2. Adler, C.P. 2000. Bones and bone tissue; normal anatomy and histology. In Bone Diseases. 1– 30. Springer-Verlag. New York.

3. Hajidakis, D., and Androulakis, I. 2006 Bone remodeling. Ann.N.Y.Acad. Sci 1092:385- 396.

4. Ducy, P., Desbois, C., Boyce, B., et al. 1996. Increased bone formation in osteocalcin- deficient mice. Nature 382: 448–452.

5. Gorsky GP. 1998 Is all bone the same? Distinctive distributions and properties on non- collagenous matrix proteis in lamellar versus woven bone imply the existence of different underlying osteogenic mechanisms. Crit Rev Oral Biol Med 9:201.

6. Zhang J, Tu Q, Chen J. 2009 Applications of transgenics in studies bone sialoprotein. J Cell Physiol 220(1):30-34.

7. Xu, T., P. Bianco, L.W. Fisher, et al. 1998. Targeted disruption of the biglycan gene leads to an osteoporosis-like phenotype in mice. Nat. Genet. 20: 78–8

8. Bianco, P., M. Riminucci, S. Gronthos, et al. 2001. Bone marrow stromal stem cells: nature, biology, and potential applications. Stem Cells 19: 180–192.

9. Ducy P., R. Zhang, V. Geoffroy, et al. 1997. Osf2/Cbfa1: a transcriptional activator of osteoblast differentiation. Cell 89: 747–754.

10. Canalis E., J. Pash, B. Gabbitas et al. 1993. Growth factors regulate the synthesis of insulin- like growth factor-I in bone cell cultures. Endocrinology 133: 33–38.

11. Rydziel, S., S. Shaikh, E. Canalis. 1994. Platelet-derived growth factor-AA and -BB (PDGF- AA and -BB) enhance the synthesis of PDGF-AA in bone cell cultures. Endocrinology 134: 2541–2546.

12. Globus, R.K., J. Plouet, D. Gospodarowicz. 1989. Cultured bovine bone cells synthesize basic fibroblast growth factor and store it in their extracellular matrix. Endocrinology 124: 1539– 1547.

13. Canalis, E., J. Pash, S. Varghese. 1993. Skeletal growth factors. Crit. Rev. Eukaryot. Gene Expr. 3: 155–166.

14. Chen, D., M. Zhao, G.R. Mundy. 2004. Bone morphogenetic proteins. Growth Factors 22: 233–241.

15. Rizzoli, R., J. Poser, U. Burgi. 1986. Nuclear thyroid hormone receptors in cultured bone cells. Metabolism 35: 71–74.

- 49- Chapter 2: Background 16. Barnard, R., K.W. Ng, T.J. Martin, et al. 1991. Growth hormone (GH) receptors in clonal osteoblast-like cells mediate a mitogenic response to GH. Endocrinology 128: 1459–1464.

17. Levy, J.R., E. Murray, S. Manolagas, et al. 1986. Demonstration of insulin receptors and modulation of alkaline phosphatase activity by insulin in rat osteoblastic cells. Endocrinology 119:1786–1792.

18. Wei, L.L., M.W. Leach, R.S. Miner, et al. 1993. Evidence for progesterone receptors in human osteoblast-like cells. Biochem. Biophys. Res. Commun. 195:525–532.

19. Clement-Lacroix P, C. Ormandy, L. Lepescheux, et al. 1999. Osteoblasts are a new target for prolactin: analysis of bone formation in prolactin receptor knockout mice. Endocrinology 140:96–105.

20. Eriksen, E.F.,D.S. Colvard, N.J.Berg, et al. 1988. Evidence of estrogen receptors in normal human osteoblast-like cells. Science 241:84–86.

21. Colvard, D.S., E.F. Eriksen, P.E. Keeting, et al. 1989. Identification of androgen receptors in normal human osteoblast-like cells. Proc. Natl. Acad. Sci. USA 86:854–857.

22. Darwish, H.M. and H.F. Deluca. 1996. Recent advances in the molecular biology of vitamin D action. Prog. Nucleic Acid Res. Mol. Biol. 53:321–344.

23. Kindmark, A., H. Torma, A. Johansson, et al. 1993. Reverse transcriptionpolymerase chain reaction assay demonstrates that the 9-cis retinoic acid receptor alpha is expressed in human osteoblasts. Biochem. Biophys. Res. Commun. 192:1367–1372.

24. Elmardi AS, Katchburian MV, Katchburian E. 1990. Electron microscopy of developing calvaria reveals images that suggest that osteoclasts engulf and destroy osteocytes during bone resorption. Calcif. Tissue Int. 46:239–245.

25. Robling A, Niziolek P, Baldridge LA, Condon KW, Allen MR, Alam I, Mantila S, Gluhak- Heinrich J, Bellido T, Harris SE and Turner CH. 2007 Mechanical stimulation of bone in vivo reduces osteocyte expression of SOST/Sclerostin. J Biol Chemistry 283(9):5866-5875.

26. Teitelbaum, S.L. 2000. Bone resorption by osteoclasts. Science 289:1504–1508.

27. Vaananen, H.K., H. Zhao, M. Mulari, et al. 2000. The cell biology of osteoclast function. J. Cell Sci. 113(Pt 3):377–381.

28. Davies J, J. Warwick, N. Totty, et al. 1989. The osteoclast functional antigen, implicated in the regulation of bone resorption, is biochemically related to the vitronectin receptor. J. Cell Biol. 109: 1817–1826.

29. Reinholt, F.P., K. Hultenby, A. Oldberg, et al. 1990. Osteopontin—a possible anchor of osteoclasts to bone. Proc. Natl. Acad. Sci. USA 87: 4473–4475.

30. Soriano, P., C. Montgomery, R. Geske, et al. 1991. Targeted disruption of the c-src proto- oncogene leads to osteopetrosis in mice. Cell 64: 693–702.

- 50- Chapter 2: Background 31. Warshawsky, H., D. Goltzman, M.F.Rouleau, et al. 1980. Direct in vivo demonstration by radioautography of specific binding sites for calcitonin in skeletal and renal tissues of the rat. J. Cell Biol. 85: 682–694.

32. Mizuno, Y., T. Hosoi, S. Inoue, et al. 1994. Immunocytochemical identification of androgen receptor in mouse osteoclast-like multinucleated cells. Calcif. Tissue Int. 54: 325–326.

33. Abu, E.O., S. Bord, A. Horner, et al. 1997. The expression of thyroid hormone receptors in human bone. Bone 21: 137–142.

34. Thomas DM, Udagawa N, Hards DK, et al. 1998. Insulin receptor expression in primary and cultured osteoclast-like cells. Bone 23: 181–186.

35. Teti, A., R. Rizzoli, Z.A. Zambonin. 1991. Parathyroid hormone binding to cultured avian osteoclasts. Biochem. Biophys. Res. Commun. 174: 1217–1222.

36. Hou, P., T. Sato, W. Hofstetter, et al. 1997. Identification and characterization of the insulin- like growth factor I receptor in mature rabbit osteoclasts. J. Bone Miner. Res. 12: 534–540.

37. Xu, L.X., T. Kukita, Y. Nakano, et al. 1996. Osteoclasts in normal and adjuvant arthritis bone tissues express the mRNA for both type I and II interleukin-1 receptors. Lab. Invest. 75: 677– 687.

38. Hofstetter, W., A. Wetterwald, M.C. Cecchini, et al. 1992. Detection of transcripts for the receptor for macrophage colony-stimulating factor, c-fms, in murine osteoclasts. Proc. Natl. Acad. Sci. USA 89: 9637–9641.

39. Zhang, Z., J.Chen, D. Jin. 1998. Platelet-derived growth factor (PDGF)-BB stimulates osteoclastic bone resorption directly: the role of receptor beta. Biochem. Biophys. Res. Commun. 251: 190–194.

40. Frost, H.M. 1990. Skeletal structural adaptations to mechanical usage (SATMU): 2. Redefining Wolff’s law: the remodeling problem. Anat. Rec. 226: 414–422.

41. Petrtyl, M., J. Hert, P. Fiala. 1996. Spatial organization of the haversian bone in man. J. Biomech. 29: 161–169.

42. Parfitt, A.M. 1994. Osteonal and hemi-osteonal remodeling: the spatial and temporal framework for signal traffic in adult human bone. J. Cell Biochem. 55:273–286.

43. Frost, H.M. 1996. Dynamics of bone remodeling. In Bone Biodynamics. H.M. Frost, Ed.: 315– 333. Littel Brown. Boston, MA.

44. Suda T, N. Takahashi, N. Udagawa, et al. 1999. Modulation of osteoclast differentiation and function by the new members of the tumor necrosis factor receptor and ligand families. Endocr. Rev. 20: 345–357.

45. Hsu, H., D.L. Lacey, C.R. Dunstan, et al. 1999. Tumor necrosis factor receptor family member RANK mediates osteoclast differentiation and activation induced by osteoprotegerin ligand. Proc. Natl. Acad. Sci. USA 96:3540–3545.

- 51- Chapter 2: Background 46. Teitelbaum SL. 2000 Bone resorption by osteoclasts. Science 289(5484):1504-1508.

47. Graham LS et al. 2009 Oxidized lipids enhance RANKL production by T lymphocytes: implications for lipid-induced bone loss. Clin Immunol. 133(2):265-275.

48. Kim, C.H., E. Takai, H. Zhou, et al. 2003. Trabecular bone response to mechanical and parathyroid hormone stimulation: the role of mechanical microenvironment. J. Bone Miner. Res. 18: 2116–2125.

49. Chapuy, M.C., M.E. Arlot, F. Duboeuf, et al. 1992. Vitamin D3 and calcium to prevent hip fractures in the elderly women. N. Engl. J. Med. 327: 1637–1642.

50. Wang J, Zhou J, Cheng C, et al. 2004. Evidence supporting dual, IGF-Iindependent and IGF-I- dependent, roles for GH in promoting longitudinal bone growth. J. Endocrinol. 180: 247–255.

51. Weinstein, R.S., R.L. Jilka, A.M. Parfitt, et al. 1998. Inhibition of osteoblastogenesis and promotion of apoptosis of osteoblasts and osteocytes by glucocorticoids. Potential mechanisms of their deleterious effects on bone. J. Clin. Invest. 102: 274–282.

52. Britto JM, Fenton AJ, Holloway WR, et al. 1994. Osteoblasts mediate thyroid hormone stimulation of osteoclastic bone resorption. Endocrinology 134:169–176.

53. Srivastava S, Toraldo G, Weitzmann MN, et al. 2001. Estrogen decreases osteoclast formation by down-regulating receptor activator of NF-kappa B ligand (RANKL)-induced JNK activation. J. Biol. Chem. 276:8836–8840.

54. Kameda, T., H. Mano, T. Yuasa, et al. 1997. Estrogen inhibits bone resorption by directly inducing apoptosis of the bone-resorbing osteoclasts. J. Exp. Med. 186:489–495.

55. Manolagas, S.C. 2000. Birth and death of bone cells: basic regulatory mechanisms and implications for the pathogenesis and treatment of osteoporosis. Endocr. Rev. 21:115–137.

56. Sato, T., H. Kawano, S. Kato, et al. 2002. Study of androgen action in bone by analysis of androgen-receptor deficient mice. J. Bone Miner. Metab. 20: 326–330.

57. Udagawa N, Takahashi N, Akatsu T, et al. 1990. Origin of osteoclasts: mature monocytes and macrophages are capable of differentiating into osteoclasts under a suitable microenvironment prepared by bone marrow-derived stromal cells. Proc. Natl. Acad. Sci. USA 87: 7260–7264.

58. Hofbauer, L.C., S. Khosla, C.R. Dunstan, et al. 2000. The roles of osteoprotegerin and osteoprotegerin ligand in the paracrine regulation of bone resorption. J. Bone Miner. Res. 15: 2–12.

59. Hofbauer, L.C.,D.L. Lacey, C.R.Dunstan, et al. 1999. Interleukin-1 beta and tumor necrosis factor-alpha, but not interleukin-6, stimulate osteoprotegerin ligand gene expression in human osteoblastic cells. Bone 25: 255–259.

60. Thirunavukkarasu K, Miles RR, D.L. Halladay, et al. 2001. Stimulation of osteoprotegerin (OPG) gene expression by transforming growth factor-beta (TGF-beta). Mapping of the OPG promoter region that mediates TGF-beta effects. J. Biol. Chem. 276: 36241-36250.

- 52- Chapter 2: Background 61. Ma, Y.L., R.L. Cain, D.L. Halladay, et al. 2001. Catabolic effects of continuous human PTH (1–38) in vivo is associated with sustained stimulation of RANKL and inhibition of osteoprotegerin and gene-associated bone formation. Endocrinology 142: 4047–4054.

62. Kitazawa, R., S. Kitazawa, S. Maeda. 1999. Promoter structure of mouse RANKL/TRANCE/OPGL/ODF gene. Biochim. Biophys. Acta 1445: 134–141.

63. Hofbauer, L.C., F. Gori, B.L. Riggs, et al. 1999. Stimulation of osteoprotegerin ligand and inhibition of osteoprotegerin production by glucocorticoids in human osteoblastic lineage cells: potential paracrine mechanisms of glucocorticoidinduced osteoporosis. Endocrinology 140: 4382–4389.

64. Saika, M., D. Inoue, S. Kido, et al. 2001. 17beta-estradiol stimulates expression of osteoprotegerin by a mouse stromal cell line, ST-2, via estrogen receptor-alpha. Endocrinology 142: 2205–2212.

65. Yano, K., N. Nakagawa, H.Yasuda, et al. 2001. Synovial cells from a patient with rheumatoid arthritis produce osteoclastogenesis inhibitor factor/osteoprotegerin: reciprocal regulation of the production by inflammatory cytokines and basic fibroblast growth factor. J. Bone Miner. Metab. 6: 365–372.

66. Liu, X.H., A. Kirschenbaum, S. Yao, et al. 2005. Cross-talk between the interleukin-6 and prostaglandin E(2) signaling systems results in enhancement of osteoclastogenesis through effects on the osteoprotegerin/receptor activator of nuclear factor-{kappa}B (RANK) ligand/RANK system. Endocrinology 146:1991–1998.

67. Moonga, B.S., O.A. Adebanjo, H.J. Wang, et al. 2002. Differential effects of interleukin-6 receptor activation on intracellular signaling and bone resorption by isolated rat osteoclasts. J. Endocrinol. 173: 395–405.

68. Sims, N.A., B.J. Jenkins, J.M. Quinn, et al. 2004. Gp130 regulates bone turnover and bone size by distinct downstream signaling pathways. J. Clin. Invest. 113:379–389.

69. Miao, D., B.He, Y. Jiang, et al. 2005. Osteoblast-derived PTHrP is a potent endogenous bone anabolic agent that modifies the therapeutic efficacy of administered PTH 1–34. J. Clin. Invest. 115: 2402–2411.

70. van Rietbergen B, Eckstein F, Koller B, Huiskes R, Baaijens F, Rüegsegger P 2000 Trabecular bone tissue strains in the healthy and osteoporotic human femur. Trans.Orthop.Res.Soc. 46:33.

71. Marcus R 1995 Relationship of age-related decreases in muscle mass and strength to skeletal status. J Gerontol A Biol Sci Med Sci 50 Spec No:86-7.

72. Layne JE, Nelson ME 1999 The effects of progressive resistance training on bone density: a review. Med Sci Sports Exerc 31(1):25-30.

73. Evans WJ 1995 Exercise, nutrition, and aging. Clin Geriatr Med 11(4):725-34.

74. Butler RN, Davis R, Lewis CB, Nelson ME, Strauss E 1998 Physical fitness: benefits of exercise for the older patient. 2. Geriatrics 53(10):46, 49-52, 61-2.

- 53- Chapter 2: Background 75. Frost HM 1987 The mechanostat: a proposed pathogenic mechanism of osteoporoses and the bone mass effects of mechanical and nonmechanical agents. Bone Miner 2(2):73-85.

76. Albright JA, Brand RA 1987 The scientific basis of orthopaedics, 2 ed. Appleton-Century Crofts, New York.

77. Rodan GA 1997 Bone mass homeostasis and bisphosphonate action. Bone 20(1):1-4.

78. Forwood MR, Turner CH 1995 Skeletal adaptations to mechanical usage: results from tibial loading studies in rats. Bone 17(4 Suppl):197S-205S.

79. Weinreb M, Rodan GA, Thompson DD 1989 Osteopenia in the immobilized rat hind limb is associated with increased bone resorption and decreased bone formation. Bone 10(3):187-94.

80. Li XJ, Jee WS, Chow SY, Woodbury DM 1990 Adaptation of cancellous bone to aging and immobilization in the rat: a single photon absorptiometry and histomorphometry study. Anat Rec 227(1):12-24.

81. Rambaut PC, Goode AW 1985 Skeletal changes during space flight. Lancet 2(8463):1050-2.

82. Minaire P 1989 Immobilization osteoporosis: a review. Clin Rheumatol 8 Suppl 2:95-103.

83. Collet P, Uebelhart D, Vico L, Moro L, Hartmann D, Roth M, Alexandre C 1997 Effects of 1- and 6-month spaceflight on bone mass and biochemistry in two humans. Bone 20(6):547-51.

84. Dehority W, Halloran BP, Bikle DD, Curren T, Kostenuik PJ, Wronski TJ, Shen Y, Rabkin B, Bouraoui A, Morey-Holton E 1999 Bone and hormonal changes induced by skeletal unloading in the mature male rat. Am J Physiol 276(1 Pt 1):E62-9.

85. Lafage-Proust MH, Collet P, Dubost JM, Laroche N, Alexandre C, Vico L 1998 Space-related bone mineral redistribution and lack of bone mass recovery after reambulation in young rats. Am J Physiol 274(2 Pt 2):R324-34.

86. Boyd SK, Ronsky JL 2000 Normal and ACL-deficient in-situ measurement of patellofemoral joint contact. J App Biomech 16:111-123.

87. Boyd SK, Matyas JR, Wohl GR, Kantzas A, Zernicke RF 2000 Early regional adaptation of periarticular bone mineral density after anterior cruciate ligament injury. J Appl Phys 89(6):2359-2364.

88. Boyd SK, Müller R, Matyas JR, Wohl GR, Zernicke RF 2000 Early morphometric and anisotropic change in periarticular cancellous bone in a model of experimental knee osteoarthritis quantified using microcomputed tomography. Clin Biomech (Bristol, Avon) 15(8):624-631.

89. Boyd SK, Muller R, Zernicke RF 2002 Mechanical and architectural bone adaptation in early stage experimental osteoarthritis. J Bone Miner Res 17(4):687-94.

90. Silbermann M, Bar-Shira-Maymon B, Coleman R, Reznick A, Weisman Y, Steinhagen- Thiessen E, von der Mark H, von der Mark K 1990 Long-term physical exercise retards trabecular bone loss in lumbar vertebrae of aging female mice. Calcif Tissue Int 46(2):80-93.

- 54- Chapter 2: Background 91. Guldberg RE, Caldwell NJ, Guo XE, Goulet RW, Hollister SJ, Goldstein SA 1997 Mechanical stimulation of tissue repair in the hydraulic bone chamber. J Bone Miner Res 12(8):1295-1302.

92. Burr DB, Schaffler MB, Yang KH, Lukoschek M, Sivaneri N, Blaha JD, Radin EL 1989 Skeletal change in response to altered strain environments: is woven bone a response to elevated strain? Bone 10(3):223-33.

93. Turner CH, Akhter MP, Raab DM, Kimmel DB, Recker RR 1991 A noninvasive, in vivo model for studying strain adaptive bone modeling. Bone 12(2):73-9.

94. Turner CH, Forwood MR, Rho JY, Yoshikawa T 1994 Mechanical loading thresholds for lamellar and woven bone formation. J Bone Miner Res 9(1):87-97.

95. Torrance AG, Mosley JR, Suswillo RF, Lanyon LE 1994 Noninvasive loading of the rat ulna in vivo induces a strain-related modeling response uncomplicated by trauma or periostal pressure. Calcif Tissue Int 54(3):241-7.

96. Chow JW, Jagger CJ, Chambers TJ 1993 Characterization of osteogenic response to mechanical stimulation in cancellous bone of rat caudal vertebrae. Am J Physiol 265(2 Pt 1):E340-7.

97. Rubin C, Turner AS, Bain S, Mallinckrodt C, McLeod K 2001 Anabolism. Low mechanical signals strengthen long bones. Nature 412(6847):603-4.

98. Jagger CJ, Chow JW, Chambers TJ 1996 Estrogen suppresses activation but enhances formation phase of osteogenic response to mechanical stimulation in rat bone. J Clin Invest 98(10):2351-7.

99. Chow JW, Fox S, Jagger CJ, Chambers TJ 1998 Role for parathyroid hormone in mechanical responsiveness of rat bone. Am J Physiol 274(1 Pt 1):E146-54.

100. Turner RT 1999 Mechanical signaling in the development of postmenopausal osteoporosis. Lupus 8(5):388-92.

101. Ma Y, Jee WS, Yuan Z, Wei W, Chen H, Pun S, Liang H, Lin C 1999 Parathyroid hormone and mechanical usage have a synergistic effect in rat tibial diaphyseal cortical bone. J Bone Miner Res 14(3):439-48.

102. Rubin C, Turner AS, Muller R, Mittra E, McLeod K, Lin W, Qin YX 2002 Quantity and quality of trabecular bone in the femur are enhanced by a strongly anabolic, noninvasive mechanical intervention. J Bone Miner Res 17(2):349-57.

103. Judex S, Boyd SK, Qin YX, Simmons T, Müller R, Rubin C 2001 Anabolic adaptations of trabecular bone to low magnitude vibrations decrease peak strain levels and result in more uniform stress under load. Ann Biomed Eng:submitted.

104. Warden SJ, Turner CH 2004 Mechanotransduction in the cortical bone is most efficient at loading frequencies of 5-10 Hz. Bone 34(2):261-70.

- 55- Chapter 2: Background 105. Robling AG, Hinant FM, Burr DB, Turner CH 2002 Improved bone structure and strength after long-term mechanical loading is greatest if loading is separated into short bouts. J Bone Miner Res 17(8):1545-54.

106. Klein-Nulend J, van der Plas A, Semeins CM, Ajubi NE, Frangos JA, Nijweide PJ, Burger EH 1995 Sensitivity of osteocytes to biomechanical stress in vitro. FASEB J 9(5):441-445.

107. Sato M, Rodan GA 1989 Bone cell shape and function. In: Stein WD, Bronner F (eds.) Cell Shape: Determinants, Regulation and Regulatory Role. Academic, San Diego, pp 329-362.

108. Skerry TM 1999 Identification of novel signaling pathways during functional adaptation of the skeleton to mechanical loading: the role of glutamate as a paracrine signaling agent in the skeleton. J Bone Miner Metab 17(1):66-70.

109. Webb CM, Zaman G, Mosley JR, Tucker RP, Lanyon LE, Mackie EJ 1997 Expression of tenascin-C in bones responding to mechanical load. J Bone Miner Res 12(1):52-8.

110. Noel LS, Champion BR, Holley CL, Simmons CJ, Morris DC, Payne JA, Lean JM, Chambers TJ, Zaman G, Lanyon LE, Suva LJ, Miller LR 1998 RoBo-1, a novel member of the urokinase plasminogen activator receptor/CD59/Ly-6/snake toxin family selectively expressed in rat bone and growth plate cartilage. J Biol Chem 273(7):3878-83.

111. Muller R, Gerber SC, Hayes WC 1998 Micro-compression: a novel technique for the nondestructive assessment of local bone failure. Technol Health Care 6(5-6):433-44.

112. Chambers TJ, Evans M, Gardner TN, Turner-Smith A, Chow JW 1993 Induction of bone formation in rat tail vertebrae by mechanical loading. Bone Miner 20(2):167-78.

113. Peake MA, Cooling LM, Magnay JL, Thomas PB, El Haj AJ 2000 Selected contribution: regulatory pathways involved in mechanical induction of c-fos gene expression in bone cells. J Appl Physiol 89(6):2498-507.

114. Lean JM, Mackay AG, Chow JW, Chambers TJ 1996 Osteocytic expression of mRNA for c- fos and IGF-I: an immediate early gene response to an osteogenic stimulus. Am J Physiol 270(6 Pt 1):E937-45.

115. Lean JM, Jagger CJ, Chambers TJ, Chow JW 1995 Increased insulin-like growth factor I mRNA expression in rat osteocytes in response to mechanical stimulation. Am J Physiol 268(2 Pt 1):E318-27.

116. Chow JW, Wilson AJ, Chambers TJ, Fox SW 1998 Mechanical loading stimulates bone formation by reactivation of bone lining cells in 13-week-old rats. J Bone Miner Res 13(11):1760-7.

117. Cowin SC, Moss-Salentijn L, Moss ML 1991 Candidates for the mechanosensory system in bone. J Biomech Eng 113(2):191-7.

118. Lanyon LE 1996 Using functional loading to influence bone mass and architecture: objectives, mechanisms, and relationship with estrogen of the mechanically adaptive process in bone. Bone 18(1 Suppl):37S-43S.

- 56- Chapter 2: Background 119. Mason DJ, Suva LJ, Genever PG, Patton AJ, Steuckle S, Hillam RA, Skerry TM 1997 Mechanically regulated expression of a neural glutamate transporter in bone: a role for excitatory amino acids as osteotropic agents? Bone 20(3):199-205.

120. Webster D, Morley Pl, van lemthe GH, and Müller R. 2008 A novel in vivo mouse model for mechanically stimulated bone adaptation – a combined experimental and computational validation study. Comput Methods Biomech Biomed Engin 11(5):435-41.

121. Parfitt AM. 1977 The cellular basis of bone turnover and bone loss: a rebuttal of the osteocytic resorption-bone flow theory. Clin Orthop Relat Res; 127:236–47.

122. Kamioka H, Honjo T, Takano-Yamamoto T. 2001 A three-dimensional distribution of osteocyte processes revealed by the combination of confocal laser scanning microscopy and differential interference contrast microscopy. Bone Feb; 28(2):145–9.

123. Lanyon LE. 1993 Osteocytes, strain detection, bone modeling and remodeling. Calcif Tissue Int 53 (Suppl 1):S102–6. S106–7.

124. Feng JQ, Ward LM, Liu S, Lu Y, Xie Y, Yuan B, Yu X, Rauch F, Davis SI, Zhang S, Rios H, Drezner MK, Quarles LD, Bonewald LF, White KE. 2006 Loss of DMP1 causes rickets and osteomalacia and identifies a role for osteocytes in mineral metabolism. Nature Genetics Oct 8.

125. Lane NE, Yao W, Balooch M, Nalla RK, Balooch G, Habelitz S, Kinney JH, Bonewald LF. 2006 Glucocorticoid-treated mice have localized changes in trabecular bone material properties and osteocyte lacunar size that are not observed in placebo-treated or estrogen- deficient mice. J Bone Miner Res 21(3):466–76.

126. Burr DB, Robling AG, Turner CH. 2002 Effects of biomechanical stress on bones in animals. Bone 30(5):781–6.

127. Ehrlich PJ, Noble BS, Jessop HL, Stevens HY, Mosley JR, Lanyon LE. 2002 The effect of in vivo mechanicalloading on estrogen receptor alpha expression in rat ulnar osteocytes. J Bone Miner Res 17(9):1646–55.

128. Skerry TM, Bitensky L, Chayen J, Lanyon LE. 1989 Early strain-related changes in enzyme activity in osteocytes following bone loading in vivo. J Bone Miner Res 4(5):783–8.

129. Raab-Cullen DM, Thiede MA, Petersen DN, Kimmel DB, Recker RR. 1994 Mechanical loading stimulates rapid changes in periosteal gene expression. Calcif Tissue Int 55(6):473-8.

130. Gluhak-Heinrich J, Ye L, Bonewald LF, Feng JQ, MacDougall M, Harris SE, Pavlin D. 2003 Mechanical loading stimulates dentin matrix protein 1 (DMP1) expression in osteocytes in vivo. J Bone Miner Res 18(5):807–17.

131. Yang W, Lu Y, Kalajzic I, Guo D, Harris MA, Gluhak-Heinrich J, Kotha S, Bonewald LF, Feng JQ, Rowe DW, Turner CH, Robling AG, Harris SE. 2005 Dentin matrix protein 1 gene cis-regulation: use in osteocytes to characterize local responses to mechanical loading in vitro and in vivo. J Biol Chem 280(21):20680–90.

- 57- Chapter 2: Background 132. Zhang K, Barragan-Adjemian C, Ye L, Kotha S, Dallas M, Lu Y, Zhao S, Harris M, Harris SE, Feng JQ, Bonewald LF. 2006 E11/gp38 selective expression in osteocytes: regulation by mechanical strain and role in dendrite elongation. Mol Cell Biol 26(12):4539–52.

133. Robling AG, Bellido T, Turner CH. 2006 Mechanical loading reduces osteocyte expression of sclerostin protein. J Bone Min Res 21 (Suppl 1):S72.

134. Bakker AD, Soejima K, Klein-Nulend J, Burger EH. 2001 The production of nitric oxide and prostaglandin E(2) by primary bone cells is shear stress dependent. J Biomech 34(5):671–7.

135. Ajubi NE, Klein-Nulend J, Alblas MJ, Burger EH, Nijweide PJ. 1999 Signal transduction pathways involved in fluid flow-induced PGE2 production by cultured osteocytes. Am J Physiol 276(1 Pt 1):E171–8.

136. Forwood MR. 1996 Inducible cyclo-oxygenase (COX-2) mediates the induction of bone formation by mechanical loading in vivo. J Bone Miner Res 11(11):1688–93.

137. Hagino H, Kuraoka M, Kameyama Y, Okano T, Teshima R. 2005 Effect of a selective agonist for prostaglandin E receptor subtype EP4 (ONO-4819) on the cortical bone response to mechanical loading. Bone 36(3):444–53.

138. Johnson ML, Picconi JL, Recker RR. 2002 The gene for high bone mass. Endocrinologist 12(5):445–53.

139. Akhter MP, Wells DJ, Short SJ, Cullen DM, Johnson ML, Haynatzki GR, Babij P, Allen KM, Yaworsky PJ, Bex F, Recker RR. 2004 Bone biomechanical properties in LRP5 mutant mice. Bone 35(1):162–9.

140. Sawakami K, Robling AG, Ai M, Pitner ND, Liu D, Warden SJ, Li J, Maye P, Rowe DW, Duncan RL, Warman ML, Turner CH. 2006 The Wnt co-receptor LRP5 is essential for skeletal mechanotransduction but not for the anabolic bone response to parathyroid hormone treatment. J Biol Chem 281(33):23698–711.

141. Kamel MA, Holladay BR, Johnson ML. 2006 Potential interaction of prostaglandin and Wnt signaling pathways mediating bone cell responses to fluid flow. J Bone Min Res 21(Suppl 1):S92.

142. Tanaka K, Yamaguchi Y, Hakeda Y. 1995 Isolated chick osteocytes stimulate formation and bone-resorbing activity of osteoclast-like cells. J Bone Miner Metab 13(2):61–70.

143. Zhao S, Zhang YK, Harris S, Ahuja SS, Bonewald LF. 2002 MLO-Y4 osteocyte-like cells support osteoclast formation and activation. J Bone Miner Res 17(11):2068–79.

144. Verborgt O, Tatton NA, Majeska RJ, Schaffler MB. 2002 Spatial distribution of Bax and Bcl-2 in osteocytes after bone fatigue: complementary roles in bone remodeling regulation? J Bone Miner Res 17(5):907–14.

145. Rubin CT, Lanyon LE. 1984 Regulation of bone formation by applied dynamic loads. J Bone Joint Surg Am 66(3):397–402.

- 58- Chapter 2: Background 146. Turner CH, Forwood MR, Otter MW. 1994 Mechanotransduction in bone: do bone cells act as sensors of fluid flow? Faseb J 8(11):875–8.

147. Robling AG, Burr DB, Turner CH. 2000 Partitioning a daily mechanical stimulus into discrete loading bouts improves the osteogenic response to loading. J Bone Miner Res 15(8):1596–602.

148. Robling AG, Hinant FM, Burr DB, Turner CH. 2002 Shorter, more frequent mechanical loading sessions enhance bone mass. Med Sci Sports Exerc 34(2):196–202.

149. Cowin SC, Moss-Salentijn L, Moss ML. 1991 Candidates for the mechanosensory system in bone. J Biomech Eng 113(2):191–7.

150. Weinbaum S, Cowin SC, Zeng Y. 1994 A model for the excitation of osteocytes by mechanical loadinginduced bone fluid shear stresses. J Biomech 27(3):339–60.

151. Wang L, Wang Y, Han Y, Henderson SC, Majeska RJ, Weinbaum S, Schaffler MB. 2005 In situ measurement of solute transport in the bone lacunar-canalicular system. Proc Natl Acad Sci USA 102(33):11911–6.

152. Han Y, Cowin SC, Schaffler MB, Weinbaum S. 2004 Mechanotransduction and strain amplification in osteocyte cell processes. Proc Natl Acad Sci USA 101(47):16689–94.

153. Veno PA, Nicolella DP, Sivakumar P, Kalajzic I, Rowe DW, Bonewald LF, Harris SE, Dallas SL. 2006 Live imaging of osteocytes within their lacunae reveals cell body and dendrite motions. J Bone Min Res 21(Suppl 1):S38–S39.

154. Salter DM, Robb JE, Wright MO. 1997 Electrophysiological responses of human bone cells to mechanical stimulation: evidence for specific integrin function in mechanotransduction. J Bone Miner Res 12(7):1133–41.

155. Hynes RO. 2002 Integrins: bidirectional, allosteric signaling machines. Cell 110(6):673–87.

156. Giancotti FG. 2000 Complexity and specificity of integrin signalling. Nat Cell Biol 2(1):E13– 4.

157. Siller-Jackson AJ, Burra S, Gu S, Harris SE, Bonewald LF, Sprague E, Jiang JX. 2006 The role of α5 integrin as a mechanosensor in the regulation of connexin 43 hemichannel release of prostaglandin in response to mechanical stress. J Bone Min Res 21(Suppl 1):S72.

158. Pavalko FM, Norvell SM, Burr DB, Turner CH, Duncan RL, Bidwell JP. 2003 A model for mechanotransduction in bone cells: the load-bearing mechanosomes. J Cell Biochem 88(1):104–12.

159. Nicolella DP, Bonewald LF, Moravits DE, Lankford J. 2005 Measurement of microstructural strain in cortical bone. Eur J Morphol 42(1–2):23–9.

160. Nicolella DP, Moravits DE, Gale AM, Bonewald LF, Lankford J. 2006 Osteocyte lacunae tissue strain in cortical bone. J Biomech 39(9):1735–43.

161. Xiao Z, Zhang S, Mahlios J, Zhou G, Magenheimer BS, Guo D, Dallas SL, Maser R, Calvet JP, Bonewald L, Quarles LD. 2006 Cilia-like structures and polycystin-1 in

- 59- Chapter 2: Background osteoblasts/osteocytes and associated abnormalities in skeletogenesis and Runx2 expression. J Biol Chem 281(41):30884–95.

162. Singla V, Reiter J. 2006 The primary cilium as the cell's antenna: signaling at a sensory organelle. Science 313(5787):629–33.

163. Malone AMD, Anderson CT, Temiyasathit S, Tang J, Tummala P, Sterns T, Jacobs CR. 2006 Primary cilia: mechanosensory organelles in bone cells. J Bone Min Res 21(Suppl 1):S39.

164. Noble BS, Peet N, Stevens HY, Brabbs A, Mosley JR, Reilly GC, Reeve J, Skerry TM, Lanyon LE. 2003 Mechanical loading: biphasic osteocyte survival and targeting of osteoclasts for bone destruction in rat cortical bone. Am J Physiol Cell Physiol 284(4):C934–43.

165. Aguirre JI, Plotkin LI, Stewart SA, Weinstein RS, Parfitt AM, Manolagas SC, Bellido T. 2006 Osteocyte apoptosis is induced by weightlessness in mice and precedes osteoclast recruitment and bone loss. J Bone Miner Res 21(4):605–15.

166. Bakker A, Klein-Nulend J, Burger E. 2004 Shear stress inhibits while disuse promotes osteocyte apoptosis. Biochem Biophys Res Commun 320(4):1163–8.

167. Plotkin LI, Mathov I, Aguirre JI, Parfitt AM, Manolagas SC, Bellido T. 2005 Mechanical stimulation prevents osteocyte apoptosis: requirement of integrins, Src kinases, and ERKs. Am J Physiol Cell Physiol 289(3):C633–43.

168. Kitase Y, Jiang JX, Johnson ML, Bonewald LF. 2006 The anti-apoptotic effects of mechanical strain on oseocytes are mediated by PGE2 and monocyte chemotactic protein-3 (MCP-3); selective protection by MCP-3 against glucocorticoid (GC), but not TNF-α induced apoptosis. J Bone Min Res 21(Suppl 1):S48.

169. Barragan-Adjemian C, Nicolella DP, Dusevich V, Dallas M, Eick D, Bonewald L. 2006 Mechanism by which MLO-A5 late osteoblast/early osteocytes mineralize in culture: similarities with mineralization of lamellar bone. Calcif Tissue Int. 79(5):340-353.

170. Franz-Odendaal TA, Hall BK, Witten PE. 2006 Buried alive: how osteoblasts become osteocytes. Dev Dyn 235(1):176–90.

171. Poole KE, van Bezooijen RL, Loveridge N, Hamersma H, Papapoulos SE, Lowik CW, Reeve J. 2005 Sclerostin is a delayed secreted product of osteocytes that inhibits bone formation. Faseb J 19(13):1842–4.

172. van Bezooijen RL, Roelen BA, Visser A, van der Wee-Pals L, de Wilt E, Karperien M, Hamersma H, Papapoulos SE, ten Dijke P, Lowik CW. 2004 Sclerostin is an osteocyte- expressed negative regulator of bone formation, but not a classical BMP antagonist. J Exp Med 199(6):805–14.

173. Li X, Zhang Y, Kang H, Liu W, Liu P, Zhang J, Harris SE, Wu D. 2005 Sclerostin binds to LRP5/6 and antagonizes canonical Wnt signaling. J Biol Chem 280(20):19883–7.

- 60- Chapter 2: Background 174. Strom TM, Francis F, Lorenz B, Boddrich A, Econs MJ, Lehrach H, Meitinger T. 1997 Pex gene deletions in Gy and Hyp mice provide mouse models for X-linked hypophosphatemia. Hum Mol Genet 6(2):165–71.

175. Rowe PS, de Zoysa PA, Dong R, Wang HR, White KE, Econs MJ, Oudet CL. 2000 MEPE, a new gene expressed in bone marrow and tumors causing osteomalacia. Genomics 67(1):54–68.

176. Liu S, Zhou J, Tang W, Jiang X, Rowe DW, Quarles LD. 2006 Pathogenic role of Fgf23 in Hyp mice. Am J Physiol Endocrinol Metab 291(1):E38–49.

177. Bonewald LF. 2006 Mechanosensation and transduction in osteocytes. Bonekey Osteovision 3(10):7-15.

178. Rosen CJ, Beamer WG, Donahue LR 2001 Defining the genetics of osteoporosis: using the mouse to understand man. Osteoporos Int 12(10):803-10.

179. Silver LM 1995 Mouse Genetics. Oxford University Press Inc., New York.

180. Nadeau JH, Taylor BA 1984 Lengths of chromosomal segments conserved since divergence of man and mouse. Proc Natl Acad Sci U S A 81(3):814-8.

181. Copeland NG, Jenkins NA, Gilbert DJ, Eppig JT, Maltais LJ, Miller JC, Dietrich WF, Weaver A, Lincoln SE, Steen RG, et al. 1993 A genetic linkage map of the mouse: current applications and future prospects. Science 262(5130):57-66.

182. Erickson RP 1989 Why isn't a mouse more like a man? Trends Genet 5(1):1-3.

183. Webster D, Wasserman E, Weber F, Bab I and Müller R. 2008 Load induced changes in trabecular and cortical bone are dose dependent in both C57Bl/6 and C3H/Hej mice. Abstracts 30th Annual Meeting ASBMR, Montreal, Canada, J. Bone Miner. Res. 23:S131.

184. Imai Y. et al. 2009 Estrogens maintain bone mass by regulating expression of genes controlling function and life span in mature osteoclasts. Ann NY Acad Science 1173:E31-E39.

185. Syed FS. 2005 Mechanisms of sex steroid effects on bone. Biochem Biophys Res Commun 328:688–696.

186. Fitzpatrick LA. 2006 Estrogen therapy for postmenopausal osteoporosis. Arq Bras Endocrinol Metabol 50:705–719.

187. Deroo BJ, Korach KS. 2006 Estrogen receptors and human disease. J Clin Invest 116:561– 570.

188. Sims NA, Sabatakos G, Chen JS, et al. 2002 Regulating DeltaFosB expression in adult Tet- Off-DeltaFosB transgenicmice alters bone formation and bone mass. Bone 30:32–39.

189. Windahl SH, Andersson G, Gustafsson GA. 2002 Elucidation of estrogen receptor function in bone with the use of mouse models. Trends Endocrinol Metab 13:195–200.

190. Vanderschueren DL, Vandenput S, Boonen A, et al. 2004 Androgens and bone. Endocr Rev 25:389–425.

- 61- Chapter 2: Background 191. Handelsman DJ. 2001 Androgen action and pharmacologic uses. In: DeGroot L, editor. Endocrinology. 4th ed. Philadelphia: W.B. Saunders; p. 2232–42.

192. de Ronde W, Pols HA, van Leeuwen JP, de Jong FH. 2003 The importance of estrogens in males. Clin Endocrinol (Oxf) 58(5):529–42.

193. Purohit A, Flanagan AM, Reed MJ. 1992 Estrogen synthesis by osteoblast cell lines. Endocrinology 131(4):2027–9.

194. Van Der Eerden BC, Van De Ven J, Lowik CW,Wit JM, Karperien M. 2002 Sex steroidmetabolism in the tibial growth plate of the rat. Endocrinology 143(10):4048–55.

195. Feix M, Wolf L, Schweikert HU. 2001 Distribution of 17beta-hydroxysteroid dehydrogenases in human osteoblast-like cells. Mol Cell Endocrinol 171(1–2):163–4.

196. Kuwano Y, Fujikawa H, Watanabe A, Shimodaira K, SekizawaA, SaitoH, et al. 1997 3betahydroxysteroid dehydrogenase activity in human osteoblast-like cells. Endocr J 44(6):847–53.

197. Colvard DS, Eriksen EF, Keeting PE, Wilson EM, Lubahn DB, French FS, et al. 1989 Identification of androgen receptors in normal human osteoblast-like cells. Proc Natl Acad Sci USA 86(3):854–7.

198. Abu EO, Horner A, Kusec V, Triffitt JT, Compston JE. 1997 The localization of androgen receptors in human bone. J Clin Endocrinol Metab 82(10):3493–7.

199. Mizuno Y, Hosoi T, Inoue S, Ikegami A, Kaneki M, Akedo Y, et al. 1994 Immunocytochemical identification of androgen receptor in mouse osteoclast-like multinucleated cells. Calcif Tissue Int 54(4):325–6.

200. Eriksen EF, Colvard DS, Berg NJ, Graham ML, Mann KG, Spelsberg TC, et al. 1988 Evidence of estrogen receptors in normal human osteoblast-like cells. Science 241(4861):84– 6.

201. Bord S, Horner A, Beavan S, Compston J. 2001 Estrogen receptors alpha and beta are differentially expressed in developing human bone. J Clin Endocrinol Metab 86(5):2309–14.

202. Karsenty G. 2006 Convergence between bone and energy homeostases: leptin regulation of bone mass. Cell Metab 4:341–348.

203. Baldock PA, Sainsbury A, Couzens M, Enriquez RF, Thomas GP, Gardiner EM, Herzog H. 2002 Hypothalamic Y2 receptors regulate bone formation. J Clin Invest 109:915–921.

204. Elefteriou F, Ahn JD, Takeda S, Starbuck M, Yang X, Liu X, Kondo H, Richards W, Bannon TW, Noda M, Clement K, Vaisse C, Karsenty G. 2005 Leptin regulation of bone resorption by the sympathetic nervous system and CART. Nature 434:514–520.

205. Sato S, Hanada R, Kimura A, Abe T, Matsumoto T, Iwasaki M, Inose H, Ida T, Mieda M, Takeuchi Y, Fukumoto S, Fujita T, Kato S, Kangawa K, Kojima M, Shinomiya K, Takeda S. 2007 Central control of bone remodeling by neuromedin U. Nat Med 13:1234–1240.

- 62- Chapter 2: Background 206. Ahima RS, Flier JS. 2000 Leptin. Annu Rev Physiol 62:413–437.

207. Ducy P, Amling M, Takeda S, Priemel M, Schilling AF, Beil FT, Shen J, Vinson C, Rueger JM, Karsenty G. 2000 Leptin inhibits bone formation through a hypothalamic relay: a central control of bone mass. Cell 100:197–207.

208. Cock TA, Back J, Elefteriou F, Karsenty G, Kastner P, Chan S, Auwerx J. 2004 Enhanced bone formation in lipodystrophic PPARgamma(hyp ⁄ hyp) mice relocates haematopoiesis to the spleen. EMBO Rep 5:1007–1012.

209. Tartaglia LA. 1997 The leptin receptor. J Biol Chem 272:6093–6096.

210. Takeda S, Elefteriou F, Levasseur R, Liu X, Zhao L, Parker KL, Armstrong D, Ducy P, Karsenty G. 2002 Leptin regulates bone formation via the sympathetic nervous system. Cell 111:305–317.

211. Flier JS. 2004 Obesity wars: molecular progress confronts an expanding epidemic. Cell 116:337–350.

212. Bray GA, York DA. 1998 The MONA LISA hypothesis in the time of leptin. Recent Prog Horm Res 53:95–117.

213. Satoh N, Ogawa Y, Katsuura G, Numata Y, Tsuji T, Hayase M, Ebihara K, Masuzaki H, Hosoda K, Yoshimasa Y, Nakao K. 1999 Sympathetic activation of leptin via the ventromedial hypothalamus: leptin-induced increase in catecholamine secretion. Diabetes 48:1787–1793.

214. Mann JJ, McBride PA, Brown RP, Linnoila M, Leon AC, DeMeo M, Mieczkowski T, Myers JE, and Stanley M. 1992 Relationship between central and peripheral serotonin indexes in depressed and suicidal psychiatric inpatients. Arch. Gen. Psychiatry 49:442–446.

215. Walther DJ, Peter JU, Bashammakh S, Hortnagl H, Voits M, Fink H, and Bader M. 2003 Synthesis of serotonin by a second tryptophan hydroxylase isoform. Science 299:76.

216. Yadav VK, Ryu JH., Suda N, Tanaka KF, Gingrich JA, Schutz G, Glorieux FH, Chiang CY, Zajac JD, Insogna KL, et al. 2008 Lrp5 controls bone formation b inhibiting serotonin synthesis in the duodenum. Cell 135:825–837.

217. Lin S, Boey D, Herzog H. 2004 NPY and Y receptors: lessons from transgenic and knockout models. Neuropeptides 38:189–200.

218. Baldock PA, Allison SJ, Lundberg P, Lee NJ, Slack K, Lin EJ, Enriquez RF, McDonald MM, Zhang L, During MJ, Little DG, Eisman JA, Gardiner EM, Yulyaningsih E, Lin S, Sainsbury A, Herzog H. 2007 Novel role of Y1 receptors in the coordinated regulation of bone and energy homeostasis. J Biol Chem 282:19092–19102.

219. Lundberg P, Allison SJ, Lee NJ, Baldock PA, Brouard N, Rost S, Enriquez RF, Sainsbury A, Lamghari M, Simmons P, Eisman JA, Gardiner EM, Herzog H. 2007 Greater bone formation of Y2 knockout mice is associated with increased osteoprogenitor numbers and altered Y1 receptor expression. J Biol Chem 282:19082–19091.

- 63- Chapter 2: Background 220. Sainsbury A, Baldock PA, Schwarzer C, Ueno N, Enriquez RF, Couzens M, Inui A, Herzog H, Gardiner EM. 2003 Synergistic effects of Y2 and Y4 receptors on adiposity and bone mass revealed in double knockout mice. Mol Cell Biol 23:5225–5233.

221. Idris AI, van ‘tHofRJ, Greig IR, Ridge SA, Baker D, Ross RA, Ralston SH. 2005 Regulation of bone mass, bone loss and osteoclast activity by cannabinoid receptors. Nat Med 11:774– 779.

222. Tam J, Ofek O, Fride E, Ledent C, Gabet Y, Muller R, Zimmer A, Mackie K, Mechoulam R, Shohami E, Bab I. 2006 Involvement of neuronal cannabinoid receptor CB1 in regulation of bone mass and bone remodeling. Mol Pharmacol 70:786–792.

223. Ofek O, Karsak M, Leclerc N, Fogel M, Frenkel B, Wright K, Tam J, Attar-Namdar M, Kram V, Shohami E, Mechoulam R, Zimmer A, Bab I. 2006 Peripheral cannabinoid receptor, CB2, regulates bone mass. Proc Natl Acad Sci USA 103:696–701.

224. Judex S, Zhong N, Squire M, Ye K, Donahue LR, Hadjiargyrou M, Rubin CT. 2005 Mechanical modulation of molecular signals which regulate anabolic and catabolicactivity in bone tissue. J Cell Biochem 94:982-994.

225. Ontiveros CA, McCabe LR. 2003 Simulated microgravity suppresses osteoblast phenotype, Runx2 levels and AP-1 transactivation. J Cell Biochem 88:427–437.

226. Nakashima K, Zhou X, Kunkel G, Zhang Z, Deng JM, Behringer RR, de Crombrugghe B. 2002 The novel zinc finger-containing transcription factor osterix is required for osteoblast differentiation and bone formation. Cell 108:17–29.

227. Sato M, Ochi T, Nakase T, Hirota S, Kitamura Y, Nomura S, Yasui N. 1999 Mechanical tension-stress induces expression of bone morphogenetic protein (BMP)-2 and BMP-4, but not BMP-6, BMP-7, and GDF-5 mRNA, during distraction osteogenesis. J Bone Miner Res 14:1084–1095.

228. Kawata A, Mikuni-Takagaki Y. 1998 Mechanotransduction in stretched osteocytes–temporal expression of immediate early and other genes. Biochem Biophys Res Commun 246:404–408.

229. Blumenfeld I, Srouji S, Peled M, Livne E. 2002. Metalloproteinases (MMPs-2, -3) are involved in TGF-beta and IGF-1-induced bone defect healing in 20-month-old female rats. Arch Gerontol Geriatr 35:59–69.

230. Moalli MR, Caldwell NJ, Patil P, Goldstein SA. 2000 An in vivo model for investigations of mechanical signal transduction in trabecular bone. J Bone Miner Res 15:1346–1353.

231. Weyts FA, Li YS, van Leeuwen J, Weinans H, Chien S.2002. ERK activation and alpha v beta 3 integrin signaling through Shc recruitment in response to mechanical stimulation in human osteoblasts. J Cell Biochem 87:85–92.

232. Pioletti DP, Muller JM, Rakotomanana LR, Corbeil JA, Wild E. 2003. Effect of micromechanical stimulations on osteoblasts: Development of a device simulating the mechanical situation at the bone-implant interface. J Biomech 36:131–135.

- 64- Chapter 2: Background 233. Rubin C, Turner AS, Bain SM, Mallinckrodt CA, McLeod K. 2001a. Anabolism: Low mechanical signals strengthen long bones. Nature 412:603–604.

234. Rubin J, Murphy TC, Zhu L, Roy EM, Nanes MS, Fan XA. 2003. Mechanical strain differentially regulates eNOS and RANKL expression via ERK1/2 MAP kinase. J Biol Chem 278:34018–34025.

235. Watanuki M, Sakai A, Sakata T, Tsurukami H, Miwa M, Uchida Y, Watanabe K, Ikeda K, Nakamura T. 2002. Role of inducible nitric oxide synthase in skeletal adaptation to acute increases in mechanical loading. J Bone Miner Res 17:1015–1025.

236. You J, Reilly GC, Zhen X, Yellowley CE, Chen Q, Donahue HJ, Jacobs CR. 2001. Osteopontin gene regulation by oscillatory fluid flow via intracellular calcium mobilization and activation of mitogen-activated protein kinase in MC3T3-E1 osteoblasts. J Biol Chem 276:13365–13371.

237. Rantakokko J, Uusitalo H, Jamsa T, Tuukkanen J, Aro HT, Vuorio E. 1999. Expression profiles of mRNAs for osteoblast and osteoclast proteins as indicators of bone loss in mouse immobilization osteopenia model. J Bone Miner Res 14:1934–1942.

238. Judex S, Donahue LR, Rubin C. 2002. Genetic predisposition to low bone mass is paralleled by an enhanced sensitivity to signals anabolic to the skeleton. FASEB J 16:1280–1282.

239. Robling AG, Hinant FM, Burr DB, Turner CH. 2002 Shorter, more requent mechanical loading sessions enhance bone mass. Med Sci Sports Exe 34(2):196-202.

240. Nordstrom P, Pettersson U, Lorentzon R 1998 Type of physical activity, muscle strength, and pubertal stage as determinants of bone mineral density and bone area in adolescent boys. J Bone Miner Res 13:1141–1148.

241. Bacabac RG, Smit TH, Mullender MG et al. 2004 Nitric oxide production by bone cells is fluid shear stress rate dependent. Biochem Biophys Res Commun 315:823–829.

242. Bacabac RG, Smit TH, Van Loon JJWA et al. 2006 Bone cell responses to high-frequency vibration stress: does the nucleus oscillate within the cytoplasm? FASEB J 20:858–864.

243. Mullender MG, Dijcks SJ, Bacabac RG et al. 2006 Release of nitric oxide, but not prostaglandin E2, by bone cells depends on fluid flow frequency. J Orthop Res 24:1170–1177.

244. Bacabac RG, Smit TH, Mullender MG et al. 2005 Initial stresskick is required for fluid shear stress-induced rate dependent activation of bone cells. Ann Biomed Eng 33:104–110.

245. Bacabac RG, Mizuno D, Schmidt CF et al. 2006 Microrheology and force traction of mechanosensitive bone cells. J Biomech 39 (Suppl. 1):S231–S232.

246. Bacabac RG, Mizuno D, Schmidt CF et al. 2008 Bone cell morphology, elasticity and mechanosensing. J Biomech 41:1590–1598.

- 65- Chapter 3: Developing a method for isolation of osteocyte RNA Chapter 3

- 66- Chapter 3: Developing a method for isolation of osteocyte RNA Developing a method for isolation of osteocyte RNA

Introduction

This was perhaps the most ambitious part of the present thesis, aimed at developing a robust method for the isolation of representative samples of total RNA derived from well-defined trabecular osteocytes of single mouse caudal vertebra. Osteocytes are usually regularly dispersed throughout the mineralized matrix and are connected to each other and cells on the bone surface through their processes, giving them the potential to recruit osteoclast precursors to stimulate bone resorption (1-5) and to regulate mesenchymal stem cell differentiation (6). Osteocytes are thus ideal cellular candidates for initiating biochemical responses culminating in tissue adaptation during bone growth and remodeling. Osteocytes are generally agreed to play a role in mechano-adaptation (7-9). Only a few osteocyte-specific proteins have been described. Recently, dentin matrix protein 1 (DMP1) has been reported to be specifically expressed only in osteocytes (10). DMP-1 is an acidic non-collagenous protein and is known to be present in the mineralized matrix of both dentin and bone (10-12). The monoclonal antibody (Mab) OB 7.3 against avian osteocytes has recently been shown to recognize PHEX (phosphate-regulating gene with homology to endopeptidases on the X chromosome) protein (13). The most compelling paradigm by which osteocytes influence the function and number of the executive cells of remodeling is symbolized by SOST/sclerostin. Osteocytes, but no other cells in bone, express sclerostin – the product of the SOST gene (14,15). As expected for an osteocyte-derived secreted protein, high levels of sclerostin are detected in the lacunar-canaliculi system (16). However, osteocytes are still poorly characterized because of their location and the lack of primary osteocyte isolation methods. Investigations of osteocyte functions are impeded by their difficult accessibility, especially from trabecular bone, which occupies the critical load-bearing sites of the skeleton. Compared to cortical bone, the spongiosa is enclosed in the cortical envelope and is substantially less accessible to cell isolation techniques, particularly in small laboratory animals such as mice. Apparently, similar considerations also apply to the assessment of mechanical load induced trabecular deformation. Hence, studies addressing load-induced cellular and molecular mechanisms in trabecular bone are rather few. Most of the published studies are based on histological and cytochemical observations of cortical bone sections (17,18). Wong and Cohn (1974) have developed a sequential enzyme digestion method to isolate various bone cell fractions from calvarial bones (19). Their main interest was to develop an isolation method for osteoblasts. A

- 67- Chapter 3: Developing a method for isolation of osteocyte RNA number of studies have subsequently been performed that suggest that osteocytes from calvaria can be isolated by a series of enzymatic digestions (19-22). Nijweide and Mulder (1986) and van der Plas and Nijweide (1992) have applied this method to isolate osteocytes from chicken calvaria. The work presented here is aimed at developing a protocol to isolate trabecular osteocytic intact RNA from a single mouse caudal vertebra (Fig. 3.1) for detailed characterization of load-induced changes in osteocyte gene expression.

A B

1 mm 0.45 mm

Fig. 3.1: Micro-computered tomography (µCT) images of mouse caudal vertebra: (A) Two-dimensional bone structure; (B) Three-dimensional bone structure. (Bab et al. Micro-tomographic Atlas of the Mouse Skeleton. 2007).

3.1. Separation of trabecular bone from caudal vertebra

The first step in this process was to physically extract the trabecular bone from the medullar cavity of the target caudal vertebra without contamination from cortical bone. This step is the most important in determining the yield of the final amount of total RNA. The method and tools employed here must therefore maximize the extraction of trabecular bone. To isolate total RNA from trabecular osteocytes, 12-week old C57BL/6 female mice (Füllinsdorf,

Switzerland) were sacrificed by CO2 inhalation. Immediately, the skin was peeled from the tail and caudal vertebrae C5-C7 separated individually. The cartilaginous ends of the vertebrae were cut off and the medullary trabecular bone (total of approximately 8 mm3), which contained bone marrow, was separated mechanically using a 21 Gauge sterile syringe needle (BD Microlance, Ireland), syringe of 1 ml volume and flushed into cold RNAlater (Ambion Inc., Austin, Texas) for RNA preservation. Then, a dental Micro-Drill with a steel burr (Hage&Meisinger GmbH, Germany) was used in order to collect the remaining trabeculae (Fig. 3.2)

- 68- Chapter 3: Developing a method for isolation of osteocyte RNA

A B

Fig. 3.2: Perforation by needle (A) a central core of medullary cavity of caudal vertebra. (B) Removal remaining trabecula by MicroDrill with a burr.

All manipulations were done in a ribonuclease (RNase) free environment to prevent RNase contamination. The accuracy of trabecular bone mechanical extraction was confirmed histologically, indicating well-defined separation of the central core of medullary cavity without contamination by cortical bone (Fig. 3.3).

Histology In order to confirm histologically the accuracy of trabecular bone extraction, caudal vertebrae were fixed in neutral buffered formalin, decalcified with 5% EDTA (pH 7.0) for 2–3 days, and embedded in paraffin. Paraffine sections 5-µm of thickness were performed using microtome for paraffine- embedded specimens. For Haematoxylin and Eosin (Sigma-Aldrich, Switzerland) staining: sections were deparaffinized and rehydrated: 3 x 3 min Xylene, 3 x 3 min 100% ethanol, 1 x 3 min 95% ethanol, 1 x 3 min 80% ethanol and 1 x 5 min deionized H2O. Haematoxylin staining: 1 x 3 min Haematoxylin, rinsing in deionized water, 1 x 5 min Tap water (to allow stain to develop), dipping 8-12 times (fast) acid ethanol (to destain), rinsing 2 x 1 min in Tap water, rinsing 1 x 2 min in deionized water. Blotting excess water from a slide holder. Eosin staining and dehydration: 1 x 30 sec Eosin, 3 x 5 min 95% ethanol, 3 x 5 min 100% ethanol (blotting excess ethanol before going into xylene), 3 x 15 min xylene. Toluidine Blue staining: 1 x 1 min in Toluidine Blue (Sigma-Aldrich, Switzerland), rinsing in deionized water for 5 min, dehydration in acetone for 3-5 min. Slides were cover-slipped using xylene-based Permount (Daigger Inc., USA).

- 69- Chapter 3: Developing a method for isolation of osteocyte RNA A

Cortical bone

Trabecular bone space

1mm

B Harvested trabeculae with bone marrow

Fig. 3.3: Histological appearance of (A) caudal vertebra following mechanical removal of central core of medullary cavity content, H&E stain. (B) Harvested central core of medullary cavity content with 125 µm trabecular bone without contamination of cortical bone, Toluidine Blue stain.

3.2. Enzymatic digestion of non-osteocytic cells Once the trabecular bone was harvested, the cell populations enriched with bone marrow and osteoblast/lining cells were extracted using collagenase A (Roche, Switzerland), enzmyme that breaks the peptide bonds in collagen, which is a key component of the animal extracellular matrix (32). The enzymatic extraction sequence of non-osteocytic cells in the trabecular bone consisted of the following steps carried out by gently shaking at 4°C: (1) Initial digest, 15 minutes, in 2 mg/ml of collagenase A in RNAlater. The volume of RNAlater was at least 50-fold greater than the tissue volume. (2) First extract of osteoblasts/lining cells (OBL1), 30 minutes digest in 3 mg/ml of collagenase A in RNAlater. (3) Second extract of osteoblasts/lining cells (OBL2), same conditions as in (2), constituted “Step 3”. At the end of each step the supernatant was collected, centrifuged for 10 min at 12000 xg at 4°C and the cell pellet re-suspended in RNAlater and kept at 4°C for further use. Some samples of trabecular bone tissue remaining after each step were fixed in phosphate buffered formalin, decalcified, and processed for histological analysis to confirm the degree of enzymatic “stripping” of the target cells (see above previous section). Following “Step 3”, the histological appearance was that of denuded trabeculae that contained histologically intact osteocytes (Fig. 3.4).

- 70- Chapter 3: Developing a method for isolation of osteocyte RNA

A B

Lacunae with Remaining bone marrow and osteocytes osteoblast/lining cells

m 125 µm 125 µm

Fig. 3.4: Histological appearance of (A) medullary central core following initial digest of bone marrow and osteoblast/lining cells. (B) Denuded trabeculae following Step 3 of sequential collagense digestions without removed osteoblast/lining cells.

3.3. RNA extraction from denuded trabeculae, Step 4 – osteocytes (OST)

The remaining fragments of trabecular bone containing osteocyte population, were pulverized by grinding for 2-3 min in liquid nitrogen using a motorized pestle. The pulverized tissue was quickly collected into 1 ml of TRIzol Reagent (Molecular Research Center, Inc., Cincinnati, OH) and homogenised for 30 seconds. Total RNA further prepared using the chloroform-phenol procedure, according to the manufactures instructions (30).

The detailed protocol for RNA extraction includes the following steps:

Homogenization

Adding 1 ml of TRI Reagent to 1-50 mg of pulverized tissue and vigorous vortexing for 30 seconds. Sample volume did not exceed 10% of the volume of TRI Reagent used for homogenization.

Phase separation

The homogenate was left for 5 minutes at room temperature to permit the complete dissociation of nucleoprotein complexes. Next, the homogenate was supplemented with 0.1 ml of Phase Separation Reagent (BCP, Molecular Research Center Inc.,) per 1 ml of TRI Reagent, the samples covered tightly and shaken vigorously for 60 seconds. After 5 minutes at room temperature and the mixture

- 71- Chapter 3: Developing a method for isolation of osteocyte RNA was centrifuged at 12,000 xg for 15 minutes at 4°C. Following centrifugation, the mixture separates into a lower red phenol-chloroform phase, interphase and the colorless upper aqueous phase. RNA remains exclusively in the aqueous phase whereas DNA and proteins are in the interphase and organic phase.

RNA precipitation

The aqueous (upper) phase was then separated and RNA precipitated by adding isopropanol (0.5 ml of isopropanol per 1 ml of TRI Reagent used for the initial homogenization). After 5 minutes at room temperature and the samples were centrifuged at 12,000 xg for 8 minutes at 4°C.

RNA wash

The supernatant was then removed and the RNA pellet washed (by vortexing) with 75% ethanol and subsequent centrifugation at 7,500 xg for 10 minutes at 4°C. Adding at least 1 ml of 75% ethanol per 1 ml TRI Reagent used for the initial homogenization.

RNA solubilization

The ethanol wash was removed and the RNA pellet briefly air-dried for 3 - 5 min. RNA was dissolved in 6 μl of RNase-free water by diethyl pyrocarbonate (DEPC) treatment. The RNA thus obtained fraction was then analyzed for quality and quantity using pico-kit of Agilent Bioanalyzer 2100 (Agilent Technologies, Foster City, CA). Specimens free of DNA and protein (260/280 light absorbance ratio of 1.6 – 1.9) were selected for RT-PCR, microarray and real-time RT-PCR analyses (Fig. 3.6).

- 72- Chapter 3: Developing a method for isolation of osteocyte RNA

Flow chart Rapid separation of caudal vertebra and chopping off its cartilaginous ends

Separation of trabecular bone into cold RNAlater

Initial digestion by Collagenase A (2 mg/ml) in RNAlater for 15 min at 4ºC

st Supernatant collection, RNA extraction 1 fraction

Second Collagenase A digestion (3 mg/ml) in RNAlater for 30 min at 4ºC

nd Supernatant collection, RNA extraction 2 fraction

Third Collagenase A digestion (3 mg/ml) in RNAlater for 30 min at 4ºC

rd Supernatant collection, RNA extraction 3 fraction

Pulverization of “stripped” trabeculae in liquid nitrogen

Collection pulverized tissue and immediate RNA extraction, 4th fraction

Fig. 3.5: Flow chart describing total RNA isolation and purification from bone cell fractions.

The RNA was prepared from the following cell fractions: 1. Initial digest, consisting mainly of bone marrow cells 2. OBL1 fraction, consisting mainly of osteoblasts and bone lining cells 3. OBL2 fraction, comprised of the remaining osteoblast/lining cells 4. OST fraction, consisting of RNA extracted from enzymatically denuded trabeculae

- 73- Chapter 3: Developing a method for isolation of osteocyte RNA

A B

C D

E

Fig. 3.6: (A) Example of analysis of intact total RNA sample with identified 18S and 28S ribosomal RNA subunits peaks and dominating bands on the electropherogram (murine spleen). (B) Analyzed an intact total RNA derived from Initial digest. (C) Analyzed extracted total RNA derived from OBL1 fraction. (D) Analyzed extracted total RNA derived from OBL2 fraction. (E) Analyzed an intact total RNA sample derived from OST fraction, with no degradation evidence.

- 74- Chapter 3: Developing a method for isolation of osteocyte RNA 3.4. Comparative marker gene mRNA expression between enzymatically isolated cell fractions and extracted RNA from denuded trabecular bone

Reverse transcription/polymerase chain reaction (RT-PCR) analysis

High quality total RNA from bone cell fractions was reverse transcribed into cDNA and amplified using Qiagen OneStep RT-PCR kit (Qiagen Inc., Valencia, CA). Omniscript and and Sensicript Reverse Transcriptases which are included in this “one-step” kit provide highly efficient and sensitive reverse transcription of RNA template quantity from 1 pg to 2 μg. HotStarTaq DNA Polymerase included in the kit provided hot-start PCR for highly specific amplification. For RT-PCR reaction 0.45 ng of RNA template was taken from each fraction. A mixture of 2.5 μl reverse transcriptase, 0.5 μl of dNTP’s mixture (400 μM of each dNTP), 0.5 μl of DNA polymerase, 10 Units RNase inhibitor (Promega, USA), 0.5 μl of each primer (0.6 μM of each primer) and RNase free water was added for a final volume of 12.5 μl. Thermal cycler conditions: reverse transcription 50°C for 30 min, initial PCR activation step 95°C for 15 min, 3-step cycling with denaturation, annealing and extension (time and temperature varied according to the primer) and final extension 72°C for 10 min. Cycling was carried out 35 times for all primers. Primer sets used were as follows:

1. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH), product size 190 base pairs, sense primer: 5'-CCT TCA TTG ACC TCA ACT AC-3', antisense primer: 5'-GGA AGG CCA TGC CAG TGA GC-3'; Denaturation: 94°C (30 sec), annealing: 58°C (30 sec), extension: 72°C (1 min).

2. Tissue non-specific alkaline phosphatase (TNSALP), 373 bp, sense primer, 5'-GCC CTC TCC AAG ACA TAT A-3', antisense: 5'-CCA TGA TCA CGT CGA TAT CC-3'; 94°C (20 sec), 60°C (30 sec); and 72°C (40 sec).

3. Osteocalcin, 371 bp, sense: 5'-CAA GTC CCA CAC AGC AGC TT-3', antisense: 5'-AAA GCC GAG CTG CCA GAG TT-3'; 94°C (30 sec), 58°C (30 sec); and 72°C (1 min).

4. Phosphate-regulating gene with homology to endopeptidases on the X chromosome (PHEX), 414 bp, sense: 5'-GCT TGA GCA AAA AGC CTG CC-3', antisense: 5'-ACC AGG GTG CCA CCA ATA AAC-3'; 94°C (1 min), 55°C (1 min) and 72°C (1 min).

- 75- Chapter 3: Developing a method for isolation of osteocyte RNA 5. Osteoblast/osteocyte factor gene (OF45), 482 bp, sense: 5'-ACT ATC CAC AAG TGG CCT CG-3', antisense: 5'-CTG TTG GCT TGC TCA GTT CC-3'; 94°C (1 min), 55°C (1 min) and 72°C (1 min).

6. GLAST-1 (384 bp) sense: 5'-TCA ATG CCC TGG GCC TCG TTG T-3'; antisense: 5'-GGG TGG CAG AAC TTG AGG AGG-3'; 94°C (30 sec), 58°C (30 sec) and 72°C (1 min).

7. DMP-1, 395 bp, sense: 5'-CGG CTG GTG GAC TCT CTA AG-3', antisense: 5'-CGG GGT CGT CGC TCT GCA TC-3'; 94°C (30 sec), 55°C (30 sec) and 72°C (1 min).

8. Mechno-growth factor (MGF), 353 bp, sense: 5'-GCT TGC TCA CCT TCA CCA GC-3', antisense: 5'-AAA TGT ACT TCC TTT CCT TCT C-3'; 94°C (30 sec), 55°C (45 sec) and 72°C (1 min).

9. SOST/Sclerostin (185 bp) sense: 5'-TCC TCC TGA GAA CAA CCA GAC-3', antisense: 5'- TGT CAG GAA GCG GGT GTA GTG-3'; 94°C (30 sec), 55°C (45 sec) and 72°C (1 min).

The oligonucleotide primer sets used crossed intron/exon boundaries so that eventual contaminations with genomic DNA would not be amplified in the amplification process or would generate amplicons of larger size. To display amplicons, aliquots of 10 μl of RT-PCR products were blotted and separated by 1% Agarose gel electrophorosis.

Immunohistochemistry

In order to confirm obtained results from gene expression profiling, protein expression analysis was performed using antibodies against alkaline phosphatase (ALP), specific to osteoblasts, and antibodies against DMP-1, specific for osteocytes. Cryosectioning of undecalcified bone: after surgical removal, caudal vertebrae were immediately snap-frozen in liquid nitrogen. Bone samples were stored in -80°C. Preceding cryosectioning, bone samples were embedded in Optimal Temperature Cutting (O.C.T) compound (Sakura, Tissue-Tek). Cryosections of 8-µm thickness were prepared using the cryomicrotome at a temperature set at - 24°C, providing an optimal operating temperature for the CTS. Adhesive tape from CryoJane® Tape-Transfer system (Instrumedics, USA) was fixed on the bone specimen, serving as an antiroll device and supporting sectioning performed with a tungsten carbide blade. The still frozen bone sections adhering to the tape were then transferred to 4x adhesive-coated slides, fixed by physical pressure. Cryosections were stored at -80°C. Before use, slides were UV-flashed (for duration of 15 seconds) to polymerize the adhesive.

- 76- Chapter 3: Developing a method for isolation of osteocyte RNA

Toluidine blue staining: Bone cryosections were immediately fixed in pre-cooled (-20°C) 70% ethanol for 2 minutes and washed in diethyl pyrocarbonate (DEPC)-treated water. The sections were stained in Toluidine Blue for 15 seconds followed by washing in DEPC-treated water and differentiating in 70% ethanol for 1 minute. Cryosections were dehydrated by increasing grades of ethanol from 70% to 100% for 1 minute each at -20°C. Dehydration was completed by xylene incubation for 1 minute at 4°C. Finally, sections were allowed to dry in a dessicator at 47°C for 2-3 minutes each. Slides were stored at -80°C and allowed to equilibrate to room temperature. Following immersion in PBS (without Ca2+/Mg2+) at room temperature, slides were placed inside a humid chamber. Sections were then incubated with 0.025% Trypsin solution and incubated for 15 minutes at room temperature. Following a rinse with PBS, blocking solution with 10% rabbit serum in PBS was applied to the sections. Sections were incubated for 1 hour at room temperature. Primary antibody solution was then added to the samples: 2 µg/ml of ALP goat anti-mouse (AbD Serotec, Düsseldorf, Germany) and 2 µg/ml of DMP-1 rabbit anti-mouse (Takara Bio, Otsu, Japan). Sections were incubated in a humid chamber at 4°C overnight. Following incubation, antibody slides were rinsed in PBS. Fluorescent secondary antibody solution was then added: Alexa488 donkey anti-goat (1:1000 dilution) (Interchim) and Alexa647 donkey anti-rabbit (1:1000 dilution) (Interchim). Sections were incubated for 45 minutes. Sections were rinsed with PBS and slides were mounted in anti-bleach mounting medium and stored at 4°C.

Results: RT-PCR and immunohistochemistry

RT-PCR analysis revealed (Fig. 3.7) the strong presence of TNS-ALP mRNA transcripts in Steps 1- 3 and weak presence of TNS-ALP mRNA in Step 4 (OST fraction). OC mRNA was identified in all steps. Also, the analysis showed the presence of PHEX, GLAST-1, OF45, DMP-1 and SOST transcripts in Steps 2 and 4; additionally PHEX, OF45 and SOST transcripts were also presented in Step 3. These results suggest that steps 2 and 3 were highly enriched with osteoblasts. In this system SOST, DMP-1, MGF and GLAST-1 are specific for osteocytes (10, 26,27,31). Their markedly increased levels in step 4, together with the very low TNS-ALP expression in this preparation indicate that the OST fraction consists mainly of RNA from trabecular osteocytes. These data demonstrate the feasibility of measuring load-induced change in gene expression in trabecular osteocytes.

- 77- Chapter 3: Developing a method for isolation of osteocyte RNA

Genes Cell fractions GAPDH (190 bp) Osteocalcin (371) TNS-ALP (373) PHEX (414) OF45 (482) GLAST-1 (384)

DMP-1 (395) Fig. 3.7: RT-PCR analysis of sequential digests from caudal vertebral trabecular bone. MGF (353) GAPDH was used as a loding control. Cycling was carried out 35 times for all SOST (185) primers.

LAD Initial OBL1 OBL2 OST digest Immunohistochemistry of bone cryosections ALP and DMP-1 protein expression has shown that ALP protein, specific for osteoblasts, was primarily expressed on the periphery of trabeculae, whereas DMP-1, specific for osteocyte population, was expressed ubiquitously throughout the extracellular matrix of trabeculae (Fig. 3.8). Immunohistochemical observations have confirmed our results of gene expression profiling of different bone cell fractions, indicating differential isolation RNA from trabecular osteoblast/lining cells and osteocytes.

- 78- Chapter 3: Developing a method for isolation of osteocyte RNA

A Trabecular bone B Intact trabecular bone

Cortical bone

Soft tissues

50 μm 25 μm

C Osteocyte, expressing DMP-1

Osteoblasts, expressing TNSALP

25 μm

Fig. 3.8: Immunohistochemical analysis of intact trabecular bone of caudal vertebra. (A) An overview of cryosection of caudal vertebra with intact trabecular bone, Toluidine Blue staining. (B) Intact trabecular bone of caudal vertebra, without staining. (C) TNSALP protein expression (green) and DMP-1 protein expression (red) in trabecular bone.

Interpretation of Results

The diversity of cell types present in bone tissue makes it difficult to assess the functional role of each cell type separately. Osteocytes in particular are barely accessible because they are confined in the calcified matrix and their isolation is problematic for ex-vivo and in-vitro studies, as they are terminally differentiated, and only limited information can be obtained by using transformed cell lines. Some reports are available giving methods for the enrichment of osteocytes from calvarial bones and their culture (22,24), but here we report, for the first time, that total RNA from well-

- 79- Chapter 3: Developing a method for isolation of osteocyte RNA defined trabecular osteocytes can also be isolated and potentially used for further functional genomics studies. Histology and gene expression profiling in the various cell fractions revealed that three main types of cells were obtained from trabecular bone after careful removal of the trabeculae and bone marrow elements. On the basis of histology, our assumption was that they represented fibroblasts (or pre-osteoblasts), osteoblasts, and osteocytes. This hypothesis was supported using OF45 primers for gene expression profiling, where OF45 was highly expressed only in osteoblast/osteocytes fractions, but not in the initial digest (mainly bone marrow cells). Further, we have shown by RT- PCR and immunohistochemical analysis for TNS-ALP and DMP-1 mRNA transcripts in sequentially isolated digests and protein expression, that osteoblasts have high TNS-ALP activity, but with their osteogenic differentiation TNS-ALP activity is decreasing (25). On the other hand, osteocytes express low levels of TNS-ALP but have high activity in DMP-1 expression (22). We concluded that the RNA we obtained in the final fraction was mainly from osteocytes. This assumption was further supported by the following observations. DMP-1, a member of the SIBLING family of acid phosphoproteins, is expressed in teeth and bone. DMP-1 mRNA and protein are highly and selectively expressed in osteocytes in which the protein is localized along dendritic processes (10). DMP-1 can thus be considered as a specific marker for osteocytes. Our results with respect to DMP-1 mRNA expression in the last fraction were in accordance with our histological data and confirmed the enrichment of osteocytes. SOST is strongly expressed in osteocytes within bone and is structurally most closely related to the DAN/cerberus family of BMP antagonists (26). In our experiments, SOST mRNA expression had a similar pattern to DMP-1, supporting the idea that the isolated cells in the last fraction were mainly osteocytes and not osteoblastic cells. Additionally, RT-PCR analysis revealed strong MGF expression only in the final fraction, the muscle insulin-like growth factor-I (IGF-I) mRNA splice variant (IGF-IEc) which has been identified in rodents. IGF-IEc or mechano growth factor (MGF) has been found to be up-regulated by exercise or muscle damage and might be a very promising target for investigation of anabolic response on bone tissue. As our purpose was to obtain RNA purely from trabecular osteocytes, we developed a method to delete as many bone marrow cells and osteoblasts as possible during seqential collagenase digestions. In conclusion, this method will allow the extraction of RNA well-defined osteoblast/lining cells populations and osteocytes from cancellous bone for subsequent detailed load-induced molecular characterization.

- 80- Chapter 3: Developing a method for isolation of osteocyte RNA Preliminary trials for developing a method for isolation of osteocyte RNA

Different methods have been tried in the preliminary studies of this thesis to obtain a sufficient amount and a high quality of RNA transcripts from trabecular osteocytes for downstream cDNA microarrays which unfortunately were unsuccessful. These approaches included: (i) using a decalcification agent such as chemical compound ethylenediaminetetraacetic acid (EDTA) combinied with an application of proteolytic enzyme cathepsin K (Calbiochem) for resorption of bone matrix (33,34); (ii) a laser-captured microdissection (LCM) technique to harvest a pure trabecular osteocyte population from undecalcified bone cryosections. Both methods did not yield sufficient qualitative and quantitative RNA extracts from trabecular osteocytes due to the degradation effect of decalcification reagent (treatment by EDTA) and time- consuming incubations during sequential digestions at 37ºC of trabecular content to an RNA integrity of 18S and 28S subunits. Additionally, the LCM approach did not reveal the presence of RNA transcripts derived from captured osteocytes due to technical problems in obtaining qualitatively intact RNA from a very low amount of starting cellular material in trabecular bone cryosections. The quality and quantity of total RNA were determined by an Agilent 2100 Bioanalyzer which showed a total degradation of extracted RNA (3.9).

Fig. 3.9: Bioanalyzer information sheet of RNA sample derived from trabecular osteocytes, with evidence for mRNA degradation.

- 81- Chapter 3: Developing a method for isolation of osteocyte RNA

Flow Chart

Rapid separation of caudal vertebra and chopping off its cartilaginous ends

Separation of trabecular bone into α-MEM full medium

Initial two digestion by Collagenase A (2 mg/ml) in RNAlater for 7 min each at 37ºC

Supernatant removal

Second Collagenase A digestion (3 mg/ml) in RNAlater for 30 min at 37ºC

Supernatant removal

Third Collagenase A digestion (3 mg/ml) in RNAlater for 30 min at 37ºC

Supernatant removal

Three cycles, each consisting of: a) demineralization by 10 mM of EDTA in RNAlater at 37ºC for 30 min b) matrix digestion with collagenase A (3 mg/ml) and cathepsin (200 nM) K in RNAlater at 37ºC for 1 hour

Collection and pooling into RNAlater, osteocytic population

RNA extraction using RNAeasy Mini kit (Qiagen Inc.)

Fig. 3.10: Flow chart describing trials of osteocytic RNA extraction by sequential collagenase digestions.

Laser Capture Microdissection experiments:

LCM was used to isolate highly pure cell populations from a heterogenous tissue section via direct visualization of the cells. There are two general classes of laser-capture microdissection: IR capture systems and ultraviolet cutting (UV) systems. In this study, a UV cutting system was used. The UV cutting system included a UV laser microdissection and catapulting (P.A.L.M. Microlaser Technologies GmbH). The principle components of laser microdisection technology are (i) visualization of the cells of interest via microscopy (ii) photo volatilization of cells surrounding a selected area and (iii) removal of cells of interest from the heterogeneous tissue section (35). Cryosectioning:

- 82- Chapter 3: Developing a method for isolation of osteocyte RNA After surgical removal, bone specimens were immediately snap-frozen in liquid nitrogen. Bone samples were stored in -80°C. Preceding cryosectioning, bone samples were embedded in Optimal Temperature Cutting (O.C.T) compound (Sakura, Tissue-Tek). Cryosections of 4 µm and 8 µm of thickness were prepared using the cryomicrotome at a temperature set at -24°C, providing an optimal operating temperature for the CTS. Adhesive tape was fixed on the bone specimen, serving as an antiroll device and supporting sectioning performed with a tungsten carbide blade. The still frozen bone sections adhering to the tape were then transferred to 4x adhesive-coated slides, fixed by physical pressure. Cryosections were stored at -80°C. Before use, slides were UV-flashed (for duration of 15 seconds) to polymerize the adhesive.

Fixation and staining: Bone cryosections were immediately fixed in precooled (-20°C) 70% ethanol for 2 minutes and washed in diethyl pyrocarbonate (DEPC)-treated water. The sections were stained in Toluidine Blue for 15 seconds and then washed in DEPC-treated water and differentiated in 70% ethanol for 1 minute. Cryosections were dehydrated by increasing grades of ethanol from 70% to 100% for 1 minute each at -20°C. Dehydration was completed by xylene incubation for 1 minute at 4°C. Finally, sections were allowed to dry in a dessicator at 47°C for 2-3 minutes each.

Total RNA isolation from microdissected cryosections: The sample was collected into a volume of Buffer RLT. The sample and Buffer RLT were transferred into a larger reaction vessel. The sample volume was adjusted to 75 µl. 20 ng of carrier RNA was added to the lysate before homogenization. The solution was then vortexed for 30 seconds. 75 µl of 70% ethanol was then added to the homogenized lysate, and mixed by pipetting. The sample was then added to an RNeasey MinElute Spin Column in a 2 ml collection tube. The tube was centrifuged and the flow-through was discarded. 10 µl DNase stock solution was added to 70 µl Buffer RDD. 350 µl Buffer RW1 was pipetted into the RNeasy MinElute Spin Column and centrifuged for 15 seconds at 10’000 xg. The flow-through was discarded. The RNeasy MinElute Spin Column was then transferred into a new collection tube and 500 µl Buffer RPE was pipetted onto the RNeasy MinElute Spin Column. The tube was centrifuged for 15 s at 10000 RPM. Five hundred µl of 80% ethanol was added to the RNeasy MinElute Spin Column. The tube was closed gently and centrifuged for 2 min at 10000 RPM to dry the silica-gel membrane. The flow- through was discarded. The RNeasy MinElute Spin Column was transferred to a new collection tube. The cap of the spin column was open and centrifuged at full speed for 5 minutes. The flow- through was discarded. Finally, the spin column was transferred to a new 1.5 ml collection tube and

- 83- Chapter 3: Developing a method for isolation of osteocyte RNA 14 µl of RNase-free water was pipetted directly onto the center of the silica-gel membrane. The tube was then centrifuged for 1 minute at maximum speed to elute. Finally, the quality and concentration of RNA was determined by an Agilent 2100 Bioanalyzer (Fig. 3.11).

Fig. 3.11: Bioanalyzer information sheet of RNA sample from laser-microdissected trabecular osteocytes, with no evidence for presence of mRNA transcripts.

Possible causes of RNA degradation It is possible that not enough material was microdissected to perform RNA isolation. One recent review on the use of LCM in the context of RT-PCR cited the efficiency to be 18% (36). Thus, in future studies more material should be microdissected to obtain a sufficient quantity of total RNA for downstream RT-PCR analysis. Trabecular bone has a large surface area (67% of bone surface) and thus is more prone to contamination by ubiquitous RNAses than cortical bone (37). Moreover, by sectioning the bone with 4- and 8 μm thick slices, the internal structure of the cells were exposed (38). These two factors may have further contributed to RNA degradation. The role of cryosection thickness should be assessed regarding the degree of RNA preservation.

- 84- Chapter 3: Developing a method for isolation of osteocyte RNA References 1. Lanyon LE. 1993 Osteocytes, strain detection, bone modeling and remodeling. Calcif Tissue Int 53:S102–106 discussion S106-7.

2. Bonewald L. Osteocytes. 2007. In: Marcus DF, Nelson DAL and Rosen CL, editors. Osteoporosis. 3rd Edition. Elsevier.

3. Kamioka H, Honjo T, Takano-Yamamoto T. 2001 A three-dimensional distribution of osteocyte processes revealed by the combination of confocal laser scanning microscopy and differential interference contrast microscopy. Bone 28:145–9.

4. Zhao S, Zhang YK, Harris S, Ahuja SS, Bonewald LF. 2002 MLO-Y4 osteocyte-like cells support osteoclast formation and activation. J Bone Miner Res 17:2068–79.

5. Baylink D, Sipe J, Wergedal J, Whittemore OJ. 1973 Vitamin D-enhanced osteocytic and osteoclastic bone resorption. Am J Physiol 224:1345–57.

6. Heino TJ, Hentunen TA, Vaananen HK. 2004 Conditioned medium from osteocytes stimulates the proliferation of bone marrow mesenchymal stem cells and their differentiation into osteoblasts. Exp Cell Res 294:458–68.

7. Burger EH, Klein-Nulend J 1999 Mechanotransduction in bone: role of the lacuno-canalicular network. FASEB J 13 Suppl:S101–S112

8. Donahue HJ et al. 2000 Gap junctions and biophysical regulation of bone cell differentiation. Bone 26:417–422.

9. Plas A van der, Aarden E, Feijen JH, Boer AH de, Wiltink A, Alblas MJ, Leij L de, Nijweide PJ 1994 Characteristics and properties of osteocytes in culture. J Bone Miner Res 9:1697–704.

10. Toyosawa S, Shintani S, Fujiwara T, Ooshima T, Sato A, Ijuhin N, Komori T 2001 Dentin matrix protein 1 is predominantly expressed in chicken and rat osteocytes but not in osteoblasts. J Bone Miner Res 16:2017–2026.

11. George A, Sabsay B, Simonian PA, Veis A 1993 Characterization of a novel dentin matrix acidic phosphoprotein. Implications for induction of biomineralization. J Biol Chem 268: 12624–12630.

12. Qin C, Brunn JC, Jones J, George A, Ramachandran A, Gorski JP, Butler WT 2001 A comparative study of sialic acid-rich proteins in rat bone and dentin. Eur J Oral Sci 109:133– 141.

13. Westbroek I, De Rooij KE, Nijweide PJ 2002 Osteocyte-specific monoclonal antibody MAb OB7.3 is directed against Phex protein. J Bone Miner Res 17:845-853

14. Balemans, W., Ebeling, M., Patel, N., Van Hul, E., Olson, P., Dioszegi, M., Lacza, C., Wuyts, W., Van Den Ende, J., Willems, P., Paes-Alves, A. F., Hill, S., Bueno, M., Ramos, F. J., Tacconi, P., Dikkers, F. G., Stratakis, C., Lindpaintner, K., Vickery, B., Foernzler, D., and Van Hul, W. 2001 Hum. Mol. Genet. 10:537–543

- 85- Chapter 3: Developing a method for isolation of osteocyte RNA 15. Balemans W, Patel N, Ebeling M, Van Hul E, Wuyts W, Lacza C, Dioszegi M, Dikkers FG, Hildering P, Willems PJ, Verheij JB, Lindpaintner K, Vickery B, Foernzler D, and Van Hul W. 2002 J. Med. Genet. 39:91–97

16. Staehling-Hampton K, Proll S, Paeper BW, Zhao L, Charmley P, Brown A, Gardner JC, Galas D, Schatzman RC, Beighton P, Papapoulos S, Hamersma H, and Brunkow ME. 2002 Am. J. Med. Genet. 110:144–152

17. Marotti G, Farneti D, Remaggi F, Tartari F 1998 Morphometric investigation on osteocytes in human auditory ossicles. Anat Anz 180:449–453.

18. Palumbo C, Ferretti M, Marotti G 2004 Osteocyte dendrogenesis in static and dynamic bone formation: an ultrastructural study. Anat Rec 278A:474–480.

19. Wong G, Cohn DV 1974 Separation of parathyroid hormone and calcitonin-sensitive cells from non-responsive bone cells. Nature 252:713–715.

20. Hefley T, Cushing J, Brand JS 1981 Enzymatic isolation of cells from bone: cytotoxic enzymes of bacterial collagenase. Am J Physiol 240:C234–C238

21. Mikuni-Takagaki Y, Kakai Y, Satoyoshi M, Kawano E, Suzuki Y, Kawase T, Saito S 1995 Matrix mineralization and the differentiation of osteocyte-like cells in culture. J Bone Miner Res 10:231–242.

22. Plas A van der, Nijweide PJ 1992 Isolation and purification of osteocytes. J Bone Miner Res 7:389–396.

23. Nijweide PJ, Mulder RJ 1986 Identification of osteocytes in osteoblast-like cell cultures using a monoclonal antibody specifically directed against osteocytes. Histochemistry 84:342–347.

24. Hefley TJ 1987 Utilization of FPLC-purified bacterial collagenase for the isolation of cells from bone. J Bone Miner Res 2:505–516.

25. McCarthy TL, Centrella M, Canalis E 1988 Further biochemical and molecular characterization of primary rat parietal bone cell cultures. J Bone Miner Res 3:401–408.

26. Winkler DG, Sutherland MK, Geoghegan JC, Yu C, Hayes T, Skonier JE, Shpektor D, Jonas M, Kovacevich BR, Staehling-Hampton K, Appleby M, Brunkow ME, Latham JA 2003 Osteocyte control of bone formation via sclerostin, a novel BMP antagonist. EMBO J 22:6267–6276.

27. Iida K, Itoh E, Kim DS, del Rincon J, Soschigano K, Kopchick J, Thorner M. 2004 Muscle mechano growth factor is preferentially induced by growth hormone in growth-hormone- deficient lit/lit mice. J Physiol 560(2):341-349.

28. Gu G, Nars M, Hentunen T, Metsikkö K, Väänänen HK 2006 Isolated osteocytes express functional gap junctions in vitro. Cell Tissue Res 323:263-267.

29. Cheng B, Kato Y, Zhao S, Luo J, Sprague E, Bonewald LF, Jiang JX 2001 PGE(2) is essential for gap junction-mediated intercellular communication between osteocyte-like MLO-Y4 cells in response to mechanical strain. Endocrinology 142:3464–3473.

- 86- Chapter 3: Developing a method for isolation of osteocyte RNA 30. Chomczynski P 1993 A reagent for the single-step simultaneous isolation of RNA, DNA and proteins from cell and tissue samples. BioTechniques, 15:532-537.

31. Skerry TM, Suva LJ. 2003 Investigation of the regulation of bone mass biomechanical loading: from quantitative cytochemistry to gene array. Cell Biochem Funct 21:223-229.

32. Tortora GJ, Funke BR and Case CL. 2007 Microbiology: an introduction brief addition. Pearson Benjamin Cummings. ISBN 03-213-960-30.

33. Harris DC. 2007 Quantitative Chemical Analysis. 7th ed., F.H. Freeman and Compagny, New York.

34. Troen BRL. 2006 The regulation of cathepsin K gene expression. Ann. N. Y. Acad. Sci. 1068: 165–72.

35. Caler WE, Carter DR. 1989 Bone creep-fatigue damage accumulation. Journal of Biomechanics 22:625-635.

36. Innis M. 1999 PCR applications protocols for functional genomics. San Diego: Academic press. 566.

37. Corwin SC. 2001 Bone mechanics Handbook. Boca Raton, FL: CRS Press.

38. Sakura. LCM for the histology technician. Histologic, 2006.

- 87- Chapter 4: Load-induced differential regulation mRNA of trabecular osteocytes Chapter 4

- 88- Chapter 4: Load-induced differential regulation mRNA of trabecular osteocytes Load-induced differential regulation mRNA of trabecular osteocytes Hypothesis to be tested: A previous Doctoral Thesis emanating from our team has reported an increase in trabecular bone formation following the administration of well-defined multiple doses of mechanical loading. It has been also shown, in the case of individual genes, that their expression in osteocytes is modulated by applying mechanical loads to bone. Hence, we hypothesized that single and multiple doses of cyclic mechanical loading induce changes in the expression of osteocyte gene clusters involved in inter- and intracellular signaling, as well as structural genes.

The technique for mRNA isolation from trabecular osteocytes developed in the previous chapter provided the basis for further studies aimed at elucidating global molecular events involved in the trabecular adaptation to different regimes of mechanical load, using cDNA microarrays. It is rationalized that a gross gain in bone density consequent to multiple load dosing represents the cumulative effect of repetitive single doses. Therefore, to substantiate the above hypothesis this study investigated how a single loading dose, as well as repeated daily doses, differentially affects global gene expression in trabecular osteocytes.

4.1. Single mechanical loading

The C57BL/6 mouse model and mechanical loading apparatus

In order to facilitate the investigation of the molecular events involved in trabecular bone formation a Caudal Vertebrae Axial compression Device (CVAD) has been developed by D. Webster et al. 2008 to mechanically stimulate the fifth caudal vertebrae (C5) of C57BL/6 (B6) female adult mice, via two stainless steal pins inserted into caudal vertebrae C4 and C6. The CVAD is able to apply an uniaxial, cyclical, compressive force to the fifth caudal vertebrae (C5) of C57BL/6 female mice via pins (0.5 mm diameter) inserted into the adjacent caudal vertebrae. A closed-loop feedback device complete with a graphical user-interface has been designed to apply a precisely controlled, cyclical, compressive load to the C5 vertebra in the mouse tail at a frequency of 10 Hz via two pins surgically inserted into C4 and C6 vertebrae (Fig. 4.1). The device is controlled via LabView 7.0 software (National Instruments) installed on a desktop computer which communicates with a servo control board (NI-7344 National instruments). Following amplification by a signal amplifier (MID-7654 National instruments), the servo board outputs a signal to a linear electro-magnetic actuator (LA25-42-000A, Bei–Kimco Magnetics). Compression of C5 is achieved

- 89- Chapter 4: Load-induced differential regulation mRNA of trabecular osteocytes by using the actuator to drive a shaft, mounted on linear bearings, connected to the distal-most pin. The proximal pin is clamped such that the only positive and negative translations along the axis of compression are permitted. The control system is closed by a load cell (13/2443 -16 TRANSMETRA haltec GmbH). As a quality control measure the feedback signal from the load cell is recorded and all force maxima and minima determined. Surgical insertion of the stainless steel pins was performed using a special pinning device, compatible with x-ray fluoroscopy. The device makes use of a V-clamp to simultaneously secure and automatically locate the cranial-caudal axis of the mouse tail. The coated pins are loaded into channels integral to the V-clamp and are manually pushed through the centers of the vertebrae, perpendicular to the cranial-caudal axis. A digital mobile C-arm (OEC Mini-View 6800, GE Medical Systems) was used to locate C4 and C6 (1,2).

Loading Device - Axis 1

Linear actuator Morphed stainless steel pins

IAM

A4

Load cell A3 Anesthetized mouse Desktop PC LabView

A2

Loading Device - Axis 2 A1 Servo control board

Servo Motor Drive

IAM

Applied mechanical Clamped pin signal a) b) Fig. 4.1: a) Overview of the dual axis Caudal Vertebra Axial Compression Device (CVAD). b) Fluoroscopic image of a mouse, graphically edited to show the location and form of the stainless steel pins once they have been

surgically inserted. The mechanical signal is applied to the distal most pin whilst the proximal-most pin is clamped (Webster et al. 2008)

Introduction

In addition to the genetic background of the mammalian species mechanical loading is one of the most important factors which regulate bone mass and shape. Although the basic form and development of bone are genetically encoded, their final mass and architecture are governed by adaptive mechanisms sensitive to the mechanical environment (3). The premise that bone cells are

- 90- Chapter 4: Load-induced differential regulation mRNA of trabecular osteocytes able to perceive and respond to mechanical forces is well accepted (4-6). This perception/response mechanism, also known as mechanotransduction, involves the conversion of a biophysical force into a biochemical response leading to changes in gene expression and cellular adaptation. Because of the inherent difficulties encountered during in vivo evaluations of the cellular, molecular, and mechanical behavior of bone, the majority of research has been conducted in vitro experiments. Although these studies have, advanced our understanding of mechanical signal transduction in bone, it has been difficult to assess whether they have simulated accurately the in vivo conditions. There also have been large variations in the response of bone cells in culture to exogenous administration of biochemical mediators (7). This lack of consistent reproducibility may be related to the absence of several factors experienced by bone cells in vivo including an appropriate osteoprogenitor cell population, blood supply, and mechanical strain tone. Where the direct effects of loading are concerned, it is likely that the cells affected are the osteocytes (8,9). These cells are distributed throughout the mineralized matrix and communicate with each other (10), and so they form an ideally located network of strain sensors, capable of providing information to the bone surface on the mechanical environment of a large region of bone. A significant body of circumstantial evidence supports this hypothesis. Osteocytes have a number of responses to loading that are consistent with such a role (11,12). An understanding of the biological pathways by which mechanical forces regulate the structure of bone qualitatively and quantitatively would provide opportunities to mimic or augment the response of bone to mechanical stimulation by pharmacological agents and may lead to novel strategies in the management of osteoporosis. However little is known about the cellular mechanisms responsible for trabecular bone adaptation, as only a few models are currently available for the elucidation of molecular mechanisms involved in load induced trabecular bone formation, using mainly rat caudal vertebrae (13). Studies using this model have been hampered by technical hurdles such as the unavailability of bone cell isolates used to investigate global gene expression and identify the set of mechanical load regulated genes. Furthermore the principal limitation imposed by the rat model is its current inaccessibility to potential genetic manipulations. To overcome this, several studies have established mouse models for the study of cortical bone adaptation and associated biochemical pathways in response to mechanical loading using C57BL/6 and C3H/Hej inbred strains (14-16). However specific genes and or combinations of genes have yet to be discovered, moreover the focus in these studies has been on cortical bone and not trabecular bone, a significant structural component which has been shown to have a more enduring sensitivity to mechanical stimulation in mature human adults than that of cortical bone (17,18).

- 91- Chapter 4: Load-induced differential regulation mRNA of trabecular osteocytes Based on published data in mice and rats, changes in vertebral trabecular osteoblast/lining cells and osteocytes occur 30-60 min to 72 hours after the administration of a single loading dose (13,19). Accordingly, the present study analyzed an early response (6 hours) in osteocyte population to load- induced changes in the mouse global gene expression, using cDNA microarrays. DNA microarrays provide a way to analyze the expression of thousands of genes at a time and to explore the activity of new genes that are being discovered. Massive data sets in this system approach can be viewed as maps that reflect the order and logic of the genetic program, rather than the physical order of genes on chromosomes (20). Methods and computational models allow the building of gene networks of cell physiology and are under continual development. Initially, cluster analysis for genome wide expression data derived from DNA microarray uses standard statistical algorithms to arrange genes according to similarity in pattern of gene expression. The output is displayed graphically, conveying the clustering and the underlying expression data simultaneously in a form intuitive for biologists. It was found that co-expression of genes of known function with poorly characterized or novel genes may provide a simple means of gaining leads regarding the function of many genes for which information is not currently available (21). We were able to successfully implement a robust technique for selective RNA extraction from trabecular osteocytes in order to investigate differential load-regulated changes in gene expression thus identifying molecular pathways of interest to futher study their role in load-stimulated bone formation in genetically modified animals.

Experimental Design

Animals: All animal protocols were approved by the Institutional Animal Care and Use Committees of ETH Zürich. Twenty eight, 8-week old C57BL/6 female mice (Füllinsdorf, Switzerland) were housed in a husbandry unit to acclimatize to their new environment. After one week mice were divided into 2 groups (0N and 8N loading groups, Fig. 4.2). Stainless steel pins (Fine Science Tools, Germany) with a diameter of 0.5 mm were inserted into the C4 and C6 mouse vertebrae of all mice, which were then given 3 weeks to recover before loading commenced. Mice of the 0N loading group formed the sham-loading group. They were only anesthetized and and connected to the CVAD which was not activated. The single load compressive dose, applied on C5 vertebra of the 8N loading group, was a sinusoidal waveform with a frequency of 10 Hz and 3000 cycles. For the pin insertion and loading procedures mice were anesthetized using an oxygen-isoflurane mixture

(Provet Medical AG). All mice were sacrificed using CO2 inhalation 6 hours after loading.

- 92- Chapter 4: Load-induced differential regulation mRNA of trabecular osteocytes

Settling period (mice 8 weeks old) Single load/Sacrifice after 6 hours Pinning & Recovery

W0 W1 W2 W3 W4

W4_0N 14 mice

W4_8N 14 mice Total 28 mice Fig. 4.2: Schematic representation of the experimental design

RNA extraction

Isolation of total RNA from trabecular osteocytes was performed from a single caudal vertebra of each mouse according to the protocol, described above in Chapter 3. Briefly, the skin was peeled from the tail and caudal vertebra C5 was separated from each mouse. The cartilaginous ends were then cut off and the medullary trabecular bone which contained bone marrow was separated mechanically using a syringe needle, a MicroDrill, and flushing with cold RNAlater. The medullary tissue was then collected by centrifugation. The digestion sequence consisted of the following steps at 4°C: Initial digestion, 15 minutes, carried out using 2 mg/ml of collagenase A in RNAlater. Step 2, 30 minutes first digest using 3 mg/ml of collagenase A in RNAlater. Step 3, second digest using the same conditions as in Step 2. The remaining trabeculae were then rapidly grinded in a mortar by pestle under liquid nitrogen on dry ice, until complete pulverization. Total RNA was extracted from the pulverized tissue, using a conventional TRI Reagent protocol and dissolved in 6 μl of RNase free water. Extracted total RNA was analyzed on quality and quantity using Agilent Bioanalyzer 2100, taking 1 μl from the RNA sample for this measurement (Fig. 4.3). A minimum 0.5 ng of total RNA was required for a single cDNA microarray Affymetrix Mouse Genome 230 chip, following the application of NuGEN Inc. (USA) pico-RNA amplification kit for preliminary mRNA amplification.

- 93- Chapter 4: Load-induced differential regulation mRNA of trabecular osteocytes

Fig. 4.3: Bioanalyzer information sheet of intact RNA sample from C5 vertebra subjected to a single loading dose, with no evidence for degradation. This RNA was further analyzed using cDNA microarray.

Microarray Experiment

Complementary RNA preparation: The quality of the isolated RNA was determined with a NanoDrop ND 1000 (NanoDrop Technologies, Delaware, USA) and a Bioanalyzer 2100 (Agilent, Waldbronn, Germany). The cDNA was prepared from total RNA using a primer mix and reverse transcriptase (RT) (WTOvation Pico System, NuGEN, 3300-12). The primers have a DNA portion that hybridizes either to the 5’ portion of the poly (A) sequence or randomly across the transcript. SPIA amplification, a linear isothermal DNA amplification process, was used to prepare single-stranded cDNA in the antisense direction of the mRNA starting material. Single-stranded cDNA quality and quantity was determined using NanoDrop ND 1000 and Bioanalyzer 2100. Fragmented and biotin- labeled single-stranded cDNA targets were generated with the FL-Ovation cDNA Biotin Module V2 (NuGEN, 4200-12).

Array hybridization: Biotin-labeled single-stranded cDNA targets (5 μg) were mixed in 220 µl of Hybridization Mix (Affymetrix Inc., P/N 900720) containing a Hybridization Controls and Control Oligonucleotide B2 (Affymetrix Inc., P/N 900454). Samples were hybridized to GeneChip® Mouse Genome 430 2.0 arrays for 18 hours at 45°C. Arrays were then washed using an Affymetrix Fluidics Station 450 FS450 0004 protocol. An Affymetrix GeneChip Scanner 3000 (Affymetrix Inc.) was used to measure the fluorescent intensity emitted by the labeled target.

- 94- Chapter 4: Load-induced differential regulation mRNA of trabecular osteocytes Data Processing and Quality Control (QC): Raw data processing was performed using the Affymetrix AGCC software. After hybridization and scanning, probe cell intensities were calculated and summarized for the respective probe sets by means of the MAS5 algorithm (Hubbell et al., 2002). To compare the expression values of the genes from chip to chip, global scaling was performed, which resulted in the normalization of the trimmed mean of each chip to target intensity (TGT value) of 500 as detailed in the statistical algorithms description document of Affymetrix (2002). Quality control measures were considered before performing the statistical analysis. These included adequate scaling factors (between 1 and 3 for all samples) and appropriate numbers of present calls calculated by application of a signed-rank call algorithm (Liu et al., 2002).

Results

This study used Affymetrix Mouse Genome 430 2.0 microarray chip to compare the gene expression profiles of 0N (control) and 8N loaded trabecular osteocyte population, with the aim being to identify altered gene expression in acute load-induced osteocytic mRNA. The mRNA levels in four biological samples from the 0N group and five samples of the 8N group were evaluated. The statistical significance of the differences between the means of the 8N and 0N gene expression values was determined using Student's t-test. The critical value for significance was chosen as p <=0.05. The microarrays analysis revealed that 28000 and 34038 probes, out of a total 45101 probe sets per chip, showed a signal in the 0N and 8N group, respectively. A total of 331 genes whose expression levels showed a 2-fold change between the 0N and 8N groups were identified when the statistical test was carried out at a p <=0.05 level of significance. Of these genes the expression of 281 was up-regulated and that of 50 was down-regulated (Tables A1, A2, Appendix). Among the genes with significantly load-regulated expression we were able to identify a group of differentially regulated genes which have known or suspected roles in bone including regulators of osteocyte, osteoblast and osteoclast metabolism and matrix proteins. These genes included IGF-1, WNT5a, IL1rn, Xiap, Asporin, STC1, Stat5a (up-regulated) and WIF1 (down-regulated). These genes encode secreted molecules (STC1, IGF-1, WNT5a, IL1rn, WIF1), transcription factors (Stat5a), intracellular signaling molecules (Xiap) and extracellular matrix molecules (Asporin).

Discussion

- 95- Chapter 4: Load-induced differential regulation mRNA of trabecular osteocytes IGF-1, an important growth factor implicated in bone formation, has been associated with both heritable peak bone mass and the response of bone to mechanical loading. In the IGF-1 – overexpressing mice used in this preliminary study, young mice (6 weeks old) show elevated bone formation and bone mass compared with wild-type litter-mates (22). Additionally, Horowitz et al. demonstrated that IGF-1 is critical for optimal skeletal growth and maintenance (23). They used a model with congenic strain 6T, which contains a QTL (quantitative trait loci) for reduced serum IGF-I donated from C3H/HeJ on a pure C57Bl/6J (B6) background. In this study a 30%-50% reduction in IGF-I expression in bone, liver, and fat of the congenic 6T mouse, as well as lower circulating IGF-I compared with control B6 were found. 6T mice also had a greater percentage body fat, but reduced serum leptin. These changes were associated with reduced cortical and trabecular bone mineral density, impaired bone formation but no change in bone resorption. Moreover, the anabolic skeletal response to intermittent parathyroid hormone (PTH) therapy was blunted in 6T compared with B6, potentially in response to greater programmed cell death in osteocytes and osteoblasts of 6T, indicating that allelic differences in IGF-I expression impact peak bone acquisition and body composition, as well as the skeletal response to PTH. IGF-1 transgenic mice also demonstrated an increase in femoral cancellous bone volume and an increase in the osteocyte lacunae occupancy, suggesting that IGF-I may extend the osteocyte life span (24). Also, Lean et al. analyzed the expression of IGF-1, during the osteogenic response of bone to mechanical stimulation, where by this growth factor was strongly expressed in osteocytes of mechanically stimulated, but not control bones, within 30 min of the osteogenic stimulus (25). IGF-I mRNA expression increased up to 6 h, was restricted to osteocytes, and was strongly suppressed by indomethacin. Although early IGF-I mRNA expression was resistant to cycloheximide, there was a degree of suppression after 6 h, raising the possibility that IGF-I expression might be prolonged by autocrine mechanisms. Thus, our study has shown that osteocytes respond to mechanical stimulation with immediate prolonged expression of IGF-I implicating osteocytes in the osteogenic response to mechanical stimulation.

Another differentially up-regulated gene, Wnt5a, the member of WNT gene family of molecules, which recently has been revealed as mediator of the adaptive response of bone to mechanical loading, where the WNT signaling pathway is responsible for a complex array of functions in maintaining bone homeostasis. The importance of Wnt5a in multiple developmental pathways is illustrated by the phenotype of Wnt5a_/_ mice, which die at birth and show many defective features such as truncated bodies, facial abnormalities and short deformed limbs (26). Wnts belong to a

- 96- Chapter 4: Load-induced differential regulation mRNA of trabecular osteocytes family of secreted glycoproteins and have been associated with the adaptative response of bone to mechanical strain. Inactivating mutations in the human low-density lipoprotein receptor-related protein 5 (LRP5) were shown to cause osteoporosis, while gain-of-function mutations in the LRP5 co-receptor increased Wnt signaling resulting in higher bone mass (27). Although evidence is accumulating that Wnts are involved in the regulation of bone mechanical adaptation, it is unknown which cells produce Wnts in response to mechanical loading. Santos and colleagues have shown that 1 h of pulsating fluid flow (0.7±0.3 Pa, 5 Hz) up-regulated mRNA expression of Wnt3a as well as the Wnt antagonist SFRP4 in MLO-Y4 osteocytes at 1 to 3 h after cessation of the fluid flow stimulus. These results suggest that osteocytes in vitro are able to respond to fluid shear stress by modulation of mRNA expression of molecules involved in Wnt signaling. The response to PFF was different in MC3T3-E1 osteoblasts, i.e., the expression of most Wnt-related genes, including Wnt5a and c-jun, was downregulated in response to PFF which underscores the specificity of the mechano- response of osteocytes in terms of Wnt expression. Mechanical loading might thus lead to Wnt production by osteocytes thereby driving the mechanical adaptation of bone (28). A function of WNT5a gene in bone formation will be broadly discussed in Chapter 4.2. In contrast, we identified that WIF-1 was down-regulated in acute load-induced trabecular osteocytes, extracellular protein which binds to WNT proteins and inhibits their activities. WIF-1 plays a role in bone biology as a negative regulator of bone mass. In situ hybridization of WIF-1 found strong expression in osteoblasts and endosteal lining cells. Microarray profiling of murine osteoblastic cells lines stimulated by BMP2 supported that WIF-1 and sFRP2 (secreted frizzled protein) were two of the most strongly up-regulated genes during terminal osteoblasts differentiation and interestingly, WIF-1 protein was observed to be expressed in vivo in trabecular, but not in cortical bone during late phase bone cell differentiation (29). Transgenic over-expression of WIF-1 decreases BMD and increases susceptibility to bone fractures in mice. WIF-1 expression was increased in mice after 4 weeks of treatment with glucocorticoids (GCs), suggesting that it may participate in the pathogenesis of the prolonged inhibition of bone formation by GCs. Thus, it appears that antagonism of WIF-1 may have therapeutic potential in osteoporosis treatment. Interleukin 1 receptor antagonist (IL1ra) appears to play an inhibitory role in osteoclastogenesis and decreases bone resorption. IL-1 is a predominant cytokine in inflammatory conditions such as osteolytic diseases, rheumatoid arthrities and is also implicated in the activation of osteoclasts. It is considered to be a candidate, in part, for initiation of increased responsiveness of large osteoclasts, to explain the pathological bone loss noted in inflammatory diseases. Trebec et al. demonstrated that IL1ra inhibited resorptive osteoclast activity by decreasing the number of nucleuses in

- 97- Chapter 4: Load-induced differential regulation mRNA of trabecular osteocytes suppressed osteoclasts (30). Thus, a role of IL1ra in the anabolic effect of load-induced bone formation is also possible. In addition, a subset of the genes with roles in osteoblast differentiation, leading to bone mineralization was identified. Particularly those that encode secreted and extracellular matrix molecules, including STC1 and asporin which also have potential osteogenic functions. Kalamajski et al. recently showed that asporin, a member of class I small leucine-rich repeat proteoglycan (SLRP), which binds to collagen type I and in the presence of asporin molecule the number of collagen nodules, and mRNA of osteoblastic markers Osterix and Runx2 were increased (31). These results suggest that asporin directly regulated hydroxiapatite formation and increases collagen mineralization. Another differentially up-regulated gene, STC1 is a mammalian homolog of the fish calcium/phosphate-regulating polypeptide whose functions are only beginning to be elucidated. Recently, it has been demonstrated that STC1 stimulates, in an autocrine/paracrine fashion, bone mineralization by increasing phosphate uptake in osteoblasts apparently via the functional activity of the sodium-dependent phosphate transporter, Pit1. Yoshiko et al. have assessed the role of STC1 on osteoblast development in fetal rat calvaria cell cultures. STC1 mRNA and protein were differentially expressed over the time course of cultures, and dexamethasone, a potent stimulator of differentiation in this model, shifted peak STC1 expression levels to earlier time. Overexpression (using recombinant human STC1) and underexpression (antisense oligonucleotides) of STC1 accelerated and retarded, respectively, osteogenic development as well as osteopontin and osteocalcin mRNA expression in mature osteoblast cultures (32). An additional observed up-regulated gene in the study was Xiap, which is a member of the inhibitor of apoptosis family of proteins (IAP). It has been recently reported (33), that Xiap is able to block glucocorticoid-induced apoptosis in osteocytes by both inhibiting caspase activity and by activating c-Jun N-terminal kinase (JNK1). Activation of JNK1 did not necessarily correlate with the ability of IAPs to inhibit caspases. XIAP and c-IAP-2 are both capable of inhibiting caspases, but transient transfection of XIAP and not c-IAP-2 in COS or 293 cells was able to activate JNK1, suggesting that XIAP anti-apoptotic properties are achieved by two separate mechanisms. Furthermore, suppression of XIAP by either siRNA or adenovirus of antisense of XIAP induced programmed cell death and inhibited Akt-stimulated cell survival in ovarian cancer cells. These data identify Xiap as a new possible mediator for osteocyte survival for improving bone strength properties. Stat5a is a known member of STAT family transcription factors. In response to cytokines and growth factors Stat5a is phosphorylated by the receptor associated kinases, and then form homo- or heterodimers that translocate to the cell nucleus where they act as a transcription activator. Previous

- 98- Chapter 4: Load-induced differential regulation mRNA of trabecular osteocytes studies have established that the effect of growth hormone (GH) in bone is consistent with the growth-promoting roles of GH and IGF-I on this target organ (34–36), and prolactin receptor (PRLR) mRNA has been identified in osteoblasts (37). Furthermore, PRLR-/- animals show reduced ossification (196) suggesting an important, uncompensatable role for PRL signaling in bone homeostasis. LeBaron et al. identified a low but specific responsiveness to PRL of chondrocytes and osteocytes of the rat femur, indicating that at least some of the effects of PRL on bone are direct (38). GH also induced low but detectable activation of Stat5a in femoral chondrocytes, whereas osteocytes showed moderate Stat5a activation in response to GH, indicating that Stat5a in osteocytes possibly plays an important function in the anabolic effect of mechanical loading. Additional studies will be needed to determine more precisely a role of Stat5a transcription factor in osteocytes and determine age-dependent differences of relevance for osteoporosis. In this study we observed small gene expression ratios in microarray analyses. Microarray analysis is able to reliably detect small (< 2-fold) changes that prove to be biologically relevant. One of them, Sclerostin/SOST osteocyte specific marker, was down-regulated close to a 2-fold change in the acute study, suggesting a hypothesis that mechanical loading reduces the expression of sclerostin protein after 24 hours (19). The insignificant effect on SOST would be explained by 6 hours being suboptimal in the presented loading regime, as the processes should include both transcriptional down-regulation as well as degradation of the existing mRNA. This study indicates the value of the microarrays approach and shows that the power of the microarray analysis method lies in its ability to detect genome-wide, coordinated, or similarly regulated differential gene expression, pointing to perturbed signaling pathways and important downstream molecular processes. In performing this study we demonstrated that acute mechanical loading induced molecular changes in trabecular osteocytes.

4.2. Repetitive mechanical loading

Introduction Following the initial study of differential gene expression the second hypothesis to be tested was that sustained cyclic mechanical loading induces sustained expression of osteocyte inter- and intracellular signaling, as well as structural genes. Chapter 4.1 investigated the basic cellular responses to mechanical loading evoked by a single loading cycle; the question however remained whether these are acute or chronic responses. The established protocol for total RNA isolation

- 99- Chapter 4: Load-induced differential regulation mRNA of trabecular osteocytes (Chapter 3) and the mechanical loading protocol used in Chapter 4.1 were therefore employed for an extended period of time. Subsequently, the results of repeated loading could be compared and contrasted with the results of single load-induced changes to elucidate the different responses between the two loading regimes. The overall purpose of this research project was to develop a proof of principle for the study of differential gene expression in a mouse model of mechanically-induced bone formation. It is proposed here to extend that proof to investigate transient versus sustained gene expression following two loading protocols: single and repetitive. A successful realization of this strategy will enable a basis for formulating further questions and developing more rigorous experimental protocols for investigating gene expression resulting from mechanical stimulation. Most importantly it would illustrate the enormous potential of this mouse model for bone adaptation research. The interpretation of the data will follow that of the two previous specific aims identifying genes of interest, based on a single loading protocol, and will be compared to genes identified in this 4-week protocol of repeated loading. Those genes present after a single load dose, but no longer expressed after the 4-week protocol, will be considered transient. The identification of these genes would raise interesting questions on the time scale of their activation throughout the duration of the chronic loading protocol; a consideration for future research. Conversely, genes expressed after the 4-week protocol, but not previously identified in Chapter 4.1, will be considered chronic, and may also lead to interesting questions for future research. A third possible result for some genes is that there will be no differences in gene expression between the two loading protocols. This finding would confirm the assumption that the anabolic adaptation of the cancellous bone is truly an additive process resulting from three single doses of 5-minute loading regimes per week. These results would have implications for the development of loading regimes to induce anabolic bone adaptation, and provide further insight into the fundamental genetic processes governing bone adaptation.

Experimental Design The mechanical loading protocol described in Chapter 4.1 was employed for a four-week period, three times loading per week, on two groups of mice treatment. These groups were a mirror of the experimental design in Chapter 4.1., so that a direct comparison could be made between the acute loading regime and the longer-term loading regime employed in this specific aim (Fig. 4.4).

- 100- Chapter 4: Load-induced differential regulation mRNA of trabecular osteocytes

Settling period (mice 8 weeks old) Sacrifice after 6 hrs Pinning/Recovery Repetitive Loading

W0 W1 W2 W3 W4 W5 W6 W7 W8

14 mice W8_0N

Total 28 mice 14 mice W8_8N

Fig. 4.4: Schematic representation of the experimental design in chronic loading.

The analysis protocol of differential gene expression was identical to the technique developed in Chapter 3 and the total osteocytic RNA was isolated 6 hours after the last mechanical load. The selection of this time-point was based on the set of results from acute loading, which were the most intriguing.

RNA extraction

Isolation of total RNA from trabecular osteocytes was performed from a single caudal vertebra of each mouse according to the protocol, described above in Chapter 3. Briefly, the skin was peeled from the tail and caudal vertebra C5 was separated from each mouse. The cartilaginous ends were then cut off and the medullary trabecular bone which contained bone marrow was separated mechanically using a syringe needle, a MicroDrill, and flushing with cold RNAlater. The medullary tissue was then collected by centrifugation. The digestion sequence consisted of the following steps at 4°C: Initial digestion, 15 minutes, carried out using 2 mg/ml of collagenase A in RNAlater. Step 2, 30 minutes first digest using 3 mg/ml of collagenase A in RNAlater. Step 3, second digest using the same conditions as in Step 2. The remaining trabeculae were then rapidly grinded in a mortar by pestle under liquid nitrogen on dry ice, until complete pulverization. Total RNA was extracted from the pulverized tissue, using a conventional TRI Reagent protocol and dissolved in 6 μl of RNase free water. Extracted total RNA was analyzed on quality and quantity using Agilent Bioanalyzer 2100, taking 1 μl from the RNA sample for this measurement (Fig. 4.5). A minimum 0.5 ng of total RNA was required for a single cDNA microarray Affymetrix Mouse Genome 230 chip, following the application of NuGEN Inc. (USA) pico-RNA amplification kit for preliminary mRNA amplification.

- 101- Chapter 4: Load-induced differential regulation mRNA of trabecular osteocytes

Fig. 4.5: Bioanalyzer information sheet of intact RNA sample from C5 vertebra subjected to repeated loading doses, with no evidence for degradation. This RNA was further analyzed using cDNA microarray.

4.2.4. Microarray experiment

Complementary RNA preparation: The quality of the isolated RNA was determined with a NanoDrop ND 1000 (NanoDrop Technologies, Delaware, USA) and a Bioanalyzer 2100 (Agilent, Waldbronn, Germany). The cDNA was prepared from total RNA using a primer mix and reverse transcriptase (RT) (WTOvation Pico System, NuGEN, 3300-12). The primers have a DNA portion that hybridizes either to the 5’ portion of the poly (A) sequence or randomly across the transcript. SPIA amplification, a linear isothermal DNA amplification process, was used to prepare single-stranded cDNA in the antisense direction of the mRNA starting material. Single-stranded cDNA quality and quantity was determined using NanoDrop ND 1000 and Bioanalyzer 2100. Fragmented and biotin- labeled single-stranded cDNA targets were generated with the FL-Ovation cDNA Biotin Module V2 (NuGEN, 4200-12).

Array hybridization: Biotin-labeled single-stranded cDNA targets (5 μg) were mixed in 220 µl of Hybridization Mix (Affymetrix Inc., P/N 900720) containing a Hybridization Controls and Control Oligonucleotide B2 (Affymetrix Inc., P/N 900454). Samples were hybridized to GeneChip® Mouse Genome 430 2.0 arrays for 18 hours at 45°C. Arrays were then washed using an Affymetrix Fluidics Station 450 FS450 0004 protocol. An Affymetrix GeneChip Scanner 3000 (Affymetrix Inc.) was used to measure the fluorescent intensity emitted by the labeled target.

Data Processing and Quality Control (QC):

- 102- Chapter 4: Load-induced differential regulation mRNA of trabecular osteocytes Raw data processing was performed using the Affymetrix AGCC software. After hybridization and scanning, probe cell intensities were calculated and summarized for the respective probe sets by means of the MAS5 algorithm (Hubbell et al., 2002). To compare the expression values of the genes from chip to chip, global scaling was performed, which resulted in the normalization of the trimmed mean of each chip to target intensity (TGT value) of 500 as detailed in the statistical algorithms description document of Affymetrix (2002). Quality control measures were considered before performing the statistical analysis. These included adequate scaling factors (between 1 and 3 for all samples) and appropriate numbers of present calls calculated by application of a signed-rank call algorithm (Liu et al., 2002).

Results

This study used Affymetrix Mouse Genome 430 2.0 microarray chips to compare the gene expression profiles of control and 8N loaded trabecular osteocyte population, with the aim of identifying altered gene expression in repeated load-induced osteocytic mRNA. The mRNA levels in three biological samples from the 0N group and four samples of the 8N group were evaluated by microarray analysis. The statistical significance of the differences between the means of the 8N and 0N gene expression values was determined using Student's t-test. The critical value for significance was chosen as P <=0.05. The microarray analysis revealed that 27997 and 36414 probes out of a total of 45101 probe sets per chip, showed a signal in the 0N and 8N group, respectively. A total of 1342 genes whose expression levels showed a 2-fold change between the 0N and 8N groups were identified when the statistical test was carried out at a P <=0.05 level of significance. Of these genes the expression of 781 was up-regulated and that of 561 was down-regulated (Tables A.3, A.4). Among the genes with significantly load-regulated expression we were able to identify a group of differentially regulated genes which have known or suspected roles in bone including regulators of osteocyte, osteoblast and osteoclast metabolism and matrix proteins. These genes included WNT5a, DMP1, Xiap, Asporin, Stat5a, Cyclin D1, alpha-actinin and RUNX2 (up-regulated). These genes encode secreted molecules (WNT5a, Cyclin D1), transcription factors (Stat5a, RUNX2), intracellular signaling molecule (Xiap) and extracellular matrix molecules (asporin, DMP-1 and alpha-actinin).

Discussion

- 103- Chapter 4: Load-induced differential regulation mRNA of trabecular osteocytes Wnt5a, a member of WNT signaling pathway, was identified up-regulated in both single and repeated load-induced studies, indicating that this gene plays an important role as a transient and chronic mediator of the adaptive response of bone to mechanical loading. It took almost two decades to obtain evidence that Wnt signaling plays a crucial role in mammalian bone homeostasis (39,40). Two reviews have broadly covered the increasing number of factors involved in the Wnt signal transduction pathways (41,42). Wnt proteins signal through canonical (β-catenin-dependent) and non-canonical mechanisms. The activity of the canonical pathway is mediated through β- catenin, which is inactivated in the absence of Wnt ligands by an oligomeric complex that consists of glycogen synthase kinase 3β (GSK-3), axin, casein kinase 1 (CK1), adenomatous polyposis coli (APC), and Disheveled. The binding of Wnt proteins to Frizzled (Fz) receptors and its coreceptor, low-density lipoprotein receptor-related protein 5 or 6 (LRP5/6) stabilizes cytoplasmic β-catenin protein, which in turn translocates to the nucleus and activates the transcription of target genes via transcription factors including lymphoid enhancer-binding factor (LEF) and T cell factors (TCF). The non-canonical Wnt pathways also require Fz receptors and, in vertebrates, have been characterized to include at least three intracellular cascades: the protein kinase C (PKC) pathway (Wnt/Ca2+), the Rho family guanosine-5’-triphosphate (GTP)-ases pathway, and the Jun N- terminal kinase (JNK) cascade. Canonical and non-canonical pathways are involved in coordinating proper bone development, formation and growth, both pre- and postnatally (43,44). The well- documented role of canonical Wnt signaling in human bone is related to loss or gain-of-function mutations in LRP5, causing osteoporosis-pseudoglioma syndrome (39), or a high bone density syndrome (40), respectively. The abnormal phenotype of high bone mass results from increased Wnt/β-catenin signaling. Canonical Wnt signaling supports osteogenic differentiation from precursor lines and stem cell lines. During in vivo bone development, canonical Wnt signaling prevents osteoblasts from differentiating into chondrocytes (45) and targets Runx2 for osteoblast differentiation (46). Postnatally, overexpression of Wnt10b in transgenic mice increases bone mass (47). Interestingly, β-catenin signaling in differentiated osteoblasts has been shown to negatively control osteoclast formation and bone resorption through an increase in osteoprotegerin production by osteoblasts (48). Along with the canonical Wnt pathway, there is now increasing evidence for noncanonical Wnt signaling pathways influencing intracellular events responsible for skeletal development and differentiation. Tu et al. (49) have reported the role of Wnt-PKC signaling in osteoblastogenesis in vitro corresponding with PKC homozygous mutant mice exhibiting a deficit in embryonic bone formation. Also, mice deficient in the G proteins αq and α11 - required for Wnt-induced PKC

- 104- Chapter 4: Load-induced differential regulation mRNA of trabecular osteocytes activation in osteoprogenitors - have bone defects in the craniofacial skeleton. The anabolic effect of parathyroid hormone (PTH) on bone formation has been in part attributed to its stimulation of non-canonical Wnt4 signaling promoting osteoprogenitor differentiation and osteoblast development, primarily through the protein kinase A pathway (50). Non-canonical Wnt4 activation of p38 mitogen-activated protein kinase (MAPK) has also been reported to enhance osteogenic differentiation of mesenchymal stem cells (isolated from human craniofacial tissues) and to promote bone formation in rodent models (51). Similarly, non-canonical Wnt5a signaling was reported to potently transdifferentiate adipoprogenitors into osteoblasts in vitro by suppressing peroxisome proliferator-activated receptor (PPAR), a key transcription factor for adipogenesis, and inducing Runx2, a key transcription factor for osteogenesis (52). Earlier in vivo data also supported the role of Wnt5a in osteoblastogenesis with the observation of decreased trabecular bone mass in the femurs of Wnt5a+/- mice (53). Loss-of-function mutation of Wnt5a in these animals resulted in truncation of the proximal skeleton and absence of distal digits. Finally, non-canonical Wnt5a has been shown to prevent the apoptosis of osteoblast progenitors and differentiated osteoblasts, comparable to anti-apoptotic effects of canonical Wnt1 and Wnt3a (54). Interestingly, the convention of two independent Wnt pathways has remained for some time, but emerging evidence suggests that the pathways are not as autonomous as originally thought. For instance, although Wnt5a is thought to primarily function though the non-canonical pathway, it can, under certain circumstances, signal through the canonical pathway. The possibility of interaction between these two pathways may explain in part the uncertainties of the role of Wnt5a in adaptive response to mechanical loading. Asporin, Xiap and Stat5a (previously broadly described in the acute-load study in Chapter 4.1.6.) were also identified as up-regulated genes in both acute and chronic loading regimes. This finding confirms the assumption that the anabolic adaptation of the cancellous bone is truly an additive process resulting from three times per week single doses of 5-minute loading regimes. These genes may have future implications as potential agents for treatment of osteoporosis and provide further insight into the fundamental genetic processes governing bone adaptation. DMP-1, extracellular matrix molecule, has been differentially up-regulated only in chronic load- induced trabecular osteocytes and may also lead to interesting questions for future research on bone adaptation to sustained mechanical loading. DMP-1, a promoter of mineralization and mineral homeostasis, increases sequentially in response to mechanical load (55). Deletion or mutation of the DMP-1 gene, which is highly expressed in embedding osteocytes and mature osteocytes, results in hypophosphatemic rickets (8). DMP-1 is highly expressed and therefore a good marker for the

- 105- Chapter 4: Load-induced differential regulation mRNA of trabecular osteocytes osteocyte lineage (56,57) and is specifically expressed along and in the canaliculi of osteocytes within the bone matrix suggesting a role for DMP1 in osteocyte function. DMP-1 is activated in a few hours in response to mechanical loading in osteocytes in the tooth movement model (58) and in the mouse ulna loading model of bone formation (8). Potential roles for DMP-1 in osteocytes have been suggested and are related to the post-translational processing and modifications of the protein as a highly phosphorylated protein and regulator of hydroxyapatite formation (59). It has been suggested that DMP-1, depending on the proteolytic processing and phosphorylation state, regulates local mineralization processes that are carried out within the lacunae and canaliculi of osteocytes in mature bone, thus keeping the lacunae and canaliculi open to allow bone fluid flow (58). Complex networks of canaliculi, containing osteocyte dendritic processes, penetrate bone; therefore, increases or decreases in canalicular volume or changes in canalicular structural integrity could alter the dynamics of fluid flow thereby altering responses of osteocytes under various physiological or pathological load conditions. A related function for DMP-1 in osteocyte biology may be to define the structural, mechanical, and material properties of the canalicular and lacunae wall. The stiffness of this wall could play an important role in detecting and transmitting mechanical signals. In addition, a runt-related transcription factor 2 (RUNX2), expressed in mature osteoblasts and early osteocytes, was also up-regulated in sustained load-induced osteocytes, confirming the hypothesis that mechanical loading increases the expression of this crucial transcription modulator for osteogenesis affecting osteoblasts differentiation and plays a fundamental role in osteoblast maturation and homeostasis. Mutations of the RUNX2 gene in humans cause cleidocranial dysplasia. Among the various stimuli that modulate Runx2 activity, mechanical loading (strain/stretching) has been revealed to be one of the most critical signals that connect Runx2 with osteoblast function and bone remodelling through mechanotransduction (60,61). Another interesting finding is that alpha-actinin was highly up-regulated (6.3 fold change) in a load-induced chronic study, suggesting that this cross-linking protein may play an important role for bone adaptation and homeostasis. Few studies have considered how the cytoskeletal composition changes in response to mechanical loads. Network polymer models of the cytoskeleton predict that recruitment of cross-linking proteins to the filament network could be an important mechanism of controlling cell stiffness, and this has been supported by several studies (62, 63). Alpha-actinin is involved in the cellular mechanoprotective response (64), and increasing the amount of alpha- actinin in the cytoskeleton is sufficient to increase the whole cell resistance to deformation. Jackson et al. showed a 29% increasing of alpha-actinin by Western blotting in mature MC-3T3-E1 cell line in vitro after mechanical loading, demonstrating that mature osteoblasts respond to fluid shear by

- 106- Chapter 4: Load-induced differential regulation mRNA of trabecular osteocytes increasing the amount of alpha-actinin that is present in the cytoskeleton. Based on the mechanical function of this protein, it is likely that this response would be sufficient to increase the whole cell stiffness, which has been observed to occur in mature osteoblasts/early osteocytes exposed to identical mechanical loads (65). Our study adds to the wealth of evidence suggesting that alpha- actinin contributes to the whole cell response to mechanical loading and may be significant for mechanical and signaling models of the cytoskeleton and further characterizes the whole cell responses to mechanical loading. This study indicates the value of the microarrays approach and shows that the power of the microarray analysis method lies in its ability to detect genome-wide, coordinated, or similarly regulated differential gene expression, pointing to perturbed signaling pathways and important downstream molecular processes. In performing this study we demonstrated that sustained mechanical loading induced molecular changes in trabecular osteocytes.

4.3. Functional genomics for identification of load-regulated pathways

Introduction

Microarrays made it possible to survey changes in the mRNA levels of genes on a genome-wide scale in a single experiment, promising an unbiased overview of changes in the transcriptome. New or adapted methods for mastering the statistical part of microarray analysis were introduced rapidly, including the still very popular significance analysis of microarrays (SAMs), (66). However, after the initial enthusiasm subsided, it became quite clear that even the statistically best-supported lists of up-regulated and down-regulated genes were most of the time as cryptic as the primary nucleotide sequence of the genome. There are two reasons for this: first of all many (if not almost all) genes serve multiple context-dependent functions. Not all changes in mRNA levels are directly connected to the experiment conducted. Therefore, it is no surprise that soon after 2001 the necessity to go beyond simple clustering and statistics was recognized (67).

One method of analyzing microarray data which is becoming popular is pathway analysis (also known as functional enrichment). This integrates the normalized array data and their annotations, such as metabolic pathways and gene ontology functional classifications. It can use various forms of currently available software for this purpose. Pathway analysis can identify more subtle changes in expression than the gene lists that result from univariate statistical analysis. Often stringent criteria are used to create these lists, for example the statistic P-value P≤0.05 and fold-change ≥2. Although

- 107- Chapter 4: Load-induced differential regulation mRNA of trabecular osteocytes genes with large changes in expression might be interesting, so might those in which there are more subtle changes, such as small, but consistent, changes in expression of a group of genes with related function. Pathway analysis is suited to detecting such trends and, as microarray technology improves, detection of lower levels of expression and smaller changes in expression is becoming feasible. These methods can also be used for different biological data, such as gene expression, and metabolomic and proteomic data, which indicate that they will be used increasingly to integrate data from these sources into metabolic networks. It is important to define the terms pathways and networks at the outset. The major tools for data analysis are considered as pathways and networks. Pathways are consecutive reaction steps, which are either biochemical transformations or sequences of signaling events, such as signal transduction. Both are static as predefined by previous studies. Networks, in contrast, are dynamic, as they are built de novo out of building blocks from binary interactions and are specific for each data set. The process of data analysis therefore consists of narrowing down the list of potentially many thousands (if not more) data points to something more interpretable. This can be achieved by using statistical analysis p-values, different scoring methods for the intersections between categories, and calculation of the relevance of the result to the data set in question (using the relative saturation of pathways and networks with data). MetaCore database (GeneGO Inc.) is a commercial package that contains more than 400 mammalian signaling and metabolic pathway maps available for mapping gene expression, proteomics, metabolic, and high content screening (HCS) data. The data generated can be exported from individual maps and clusters of maps and analyzed further with networks. MetaCore is a web- based computational platform for multiple applications in systems biology. It is primarily designed for the analysis of high-throughput molecular data (microarray-based and serial analysis of gene expression (SAGE) gene expression, array-comparative genomic-hybridization DNA arrays, proteomics data, metabolic profiles, and so on) in the context of human and mammalian networks, canonical pathways, diseases, and cellular processes. MetaCore is an integrated system, which consists of (1) a curated database of mammalian biology, (2) a suite of tools for querying, visualization, and statistical analysis including pathways maps, network algorithms, and filters, (3) a toolkit (pathway editor) for custom assembly of functional networks, and (4) a set of parsers for uploading and manipulating different types of high-throughput molecular data (68-70). As a foundation, MetaCore has a database of protein–protein, protein–DNA, and protein–compound interactions, metabolic reactions, pathway maps, bioactive compounds (metabolites, drugs, and ligands), and diseases. Human pathways have been manually collected from the experimental

- 108- Chapter 4: Load-induced differential regulation mRNA of trabecular osteocytes literature for more than 5 years. This represents one of the most comprehensive databases in the field, the core of MetaCore consists of more than 4.5 million individual findings resulting in about 50,000 signaling interactions and 20,000 human metabolic transformations (covering both endogenous and xenobiotic metabolism). The database has interaction information for more than 90% of known human proteins, including 1720 transcription factors and 650 GPCRs. This content is linked to 3200 human diseases and conditions. The bioactive chemistry component includes more than 7000 known drugs with protein targets and 5000 endogenous metabolites. The pathway information is organized in more than 400 signaling and metabolic maps with more than 3000 canonical pathways represented. The MetaCore software currently runs on an Intel-based 32-bit server running RedHat Linux Enterprise 3 AS (RedHat, Raleigh, NC) and the web server runs Apache 1.3.x/mod_perl. Software on the server side is written in Perl, whereas the client side requires HTML/JavaScript and the Macromedia Flash Player Plug-in (Macromedia Inc, San Francisco, CA). The MetaCore database is generated from manual annotation of full text articles as well as disease relevant information from OMIM and EntrezGene. Every node on the network is associated with genes and proteins through the tables in the general database schema. The novel database architecture enables mapping of the high-throughput experimental data associated with genes and proteins onto the networks. Every experimental data point (e.g., a set of probes on the microarray or a frequency for a certain SAGE tag) represents an attribute of the unique gene or protein identifier. Therefore, the high-throughput data can be linked with the corresponding node in the database and visualized on the networks containing this node. Visually, the altered expression or protein abundance data is presented as a solid circle above the node (red and blue represent increased and decreased abundance, respectively with a number 1 for acute and number 2 for repeated loading). The applications of this quite straightforward procedure are ubiquitous in basic research and in drug discovery. For instance, one can directly compare the lists of genes derived from different types of high-throughput or “small-scale experiments” on the same networks. When the same data type and experimental platform is used, the conditional networks can be readily compared for common and different sub-networks and patterns. Such fine- grained mapping can also be performed to compare the tissue and cell type specific response, different time-points, drug dosage, and different patients from the same cohort, and etc.

Materials and Methods

- 109- Chapter 4: Load-induced differential regulation mRNA of trabecular osteocytes In both the acute and chronic mechanical loading studies, MetaCore (GeneGo, St Joseph, MI) was used to map the differentially expressed genes into biological networks and for the functional interpretation of the experimental microarray data. MetaCore is an integrated software suite based on a manually curated database of mammalian protein-protein interactions, protein-DNA interactions, transcriptional factors, metabolic, and signaling pathways. Within MetaCore, the networks were generated as a combination of binary single-step interactions (edges) which connected proteins and genes (nodes). The nodes and edges were derived from the corresponding interaction tables in the MetaCore database and were visualized as clusters of interconnected nodes with the Macromedia Flash Player Plug-in. The end nodes on the networks had only one edge; the internal nodes had anywhere between two to several hundred edges depending on connectivity with other nodes. The networks were built from an input list of genes, corresponding to the components (network classes) in the database. The nodes in the input list were therefore considered as root nodes. The list of genes was imported as a text in an Excel file directly from Affymetrix microarray analysis software (www.affymetrix.com) and uploaded as their Swiss-Prot IDs to MetaCore, for identification of load-regulated signaling pathways. Before building networks, the interactions were preselected based on the level of trust, interaction direction, effects, mechanisms, and tissue specificity (in which only the edges with both nodes belonging to a chosen tissue remain). The nodes from the input list with no connections with other nodes on the list were removed. The edges of networks were assigned with weights depending on the type of interactions. The biological process enrichment was analyzed based on GO Ontology processes. The direct interactions algorithm was the most stringent, the only edges allowed being those between two nodes which are root nodes, e.g., objects from the list directly connected to each other. For network analysis, the shortest path algorithm was used to map the shortest path for interaction, based on Dijekstra algorithm (94).

Results To identify potential signaling pathways associated with the skeletal anabolic response to mechanical loading, we analyzed our microarray expression data using MetaCore software. This software comes with a built-in natural language processing module MedScan and a comprehensive database containing more than 150,000 events of regulation, interaction and modification between proteins and cell processes obtained from PubMed which allows it to generate a biological association network (BAN) of known biological interactions. In order to characterize change in

- 110- Chapter 4: Load-induced differential regulation mRNA of trabecular osteocytes signaling pathways involved in the response of trabecular osteocytes to mechanical loading, we imported global microarray data into MetaCore software. Functional genomics analysis revealed overall 65 load-regulated signaling pathways in single mechanical loading and 153 load-regulated signaling pathways in repeated loading study between loaded and unloaded samples (Tables A5 and A6, Appendix). The prediction with MetaCore based on a whole-genome gene expression profile, highlighted in a single loading top ten significant load-regulated pathways in trabecular osteocytes in response to mechanical loading. This includes a calcium regulated α-1A adrenergic receptor-dependent inhibition of PI3K-Act of cytoskeleton remodeling with an impact factor of 67% (meaning that 8 objects of the pathway from total 12 were differentially regulated with significance of P < 0.05), signaling pathway for regulation of translation initiation (60%), signal transduction by protein kinase A (PKA) signaling (39%), TGF-β receptor signaling development process (29%) and IGF-1 receptor 1 signaling development pathway (27%). It has been found that these genes are involved in a number of signaling pathways and have been implicated in regulating the formation and/or activity of bone cells in response to mechanical strain (71-74). In the same manner, MetaCore highlighted top ten significant load-regulated signaling pathways in repetitive loading. This includes cell adhesion of endothelial cell contacts by non-junctional mechanisms (79%), calcium regulated α-1A adrenergic receptor-dependent inhibition of PI3K-Act of cytoskeleton remodeling (75%), cytoskeleton remodeling of integrin outside-in signaling pathway (61%) growth factor activated extracellular signal-regulated kinases (ERK), (56%), cytoskeleton remodeling regulation of actin cytoskeleton by Rho GTPases (52%), cytoskeleton remodeling activation of protein kinase C (PKC) via G-protein coupled receptor (45.45%) and development of Wnt5a signaling pathway (40%, Fig. 4.6 – 4.7). It has been demonstrated that Wnt5a signaling influencing intracellular events responsible for skeletal development and differentiation (49).

- 111- Chapter 4: Load-induced differential regulation mRNA of trabecular osteocytes

Fig. 4.6: Development of WNT5a signaling pathway map (MetaCore GeneGO). Visually, the altered gene expression or protein abundance data is presented as a solid circle above the node (red and blue represent increased and decreased abundance; with a number 1 for acute and number 2 for repeated loading, respectively).

- 112- Chapter 4: Load-induced differential regulation mRNA of trabecular osteocytes

Fig. 4.7: Legends of signaling pathway map.

Additionally, we compared significantly load-regulated signaling pathways according to the biological functions between single and repetitive doses of cyclic mechanical loading (Fig. 4.8). Interestingly, percentage of load-regulated signaling pathways in osteocytes associated with immune response in repetitive loading was decreased comparing to a single loading, whereas percentage of load-regulated pathways associated with cell growth/differentiation was increased in repetitive loading comparing to a single loading. Also, percentage of signaling pathways associated with apoptosis was increased in repetitive loading comparing to a single loading. We assume that bone during repetitive loading accommodates with a time to sustained mechanical force, whereas bone response mainly by inflammation processing following single mechanical stimulation.

- 113- Chapter 4: Load-induced differential regulation mRNA of trabecular osteocytes

Biological processes A 1.5% 4.6% 4.6% Immune response G-protein signaling 6.2% 28% Cell growth/differentiation Chemotaxis

6.2% Cell cycle Translation Signal transduction 6.2% Cell death 1.5% Transcription Transport 7.7% Cytoskeleton remodeling Neurophysiological process 1.5% 20.0% Unknown 9.2% 1.5%

B Biological processes Immune response Proteolysis 3.9% 3.3% 9.8% Cell growth/differentiation 0.7% Blood coagulation 7.8% G-protein signaling Oxidative stress 3.9% Cell adhesion

4.6% Cell cycle Transcription 3.3% 29.4% Cell death Transport 5.2% Signal transduction Neurophysiological process 2.6% Translation 5.2% Cytoskeleton remodeling 0.7% 2.6% Lipid metabolism 6.5% 1.3% 5.9% Unknown

Fig. 4.8: Distribution of load-regulated signaling pathways according to cell processes in response to (A) single and (B) repetitive mechanical loading.

Discussion

The current study presents for the first time murine global-genome mRNA expression profiles in trabecular osteocytes in response to acute and chronic mechanical loading in vivo. Whole-genome microarrays analyses predicted that signaling pathways such as PI3K, ECM-receptor interactions, TGF-β signaling, and Wnt signaling are involved in trabecular osteocyte loading-driven responses

- 114- Chapter 4: Load-induced differential regulation mRNA of trabecular osteocytes in acute and chronic regimes at 6 h after the last loading. First, PI3K signaling pathway is known to be activated in response to various extracellular signals such as peptide growth factors, insulin, and insulin-like growth factors (75). Insulin-like growth factors 1 and 2, for instance, can stimulate bone formation due to mechanical load (76-78). Signaling pathways downstream of PI3K affect a wide range of cellular activities including cell growth, cell survival, and cell movement (79). Second, mRNA levels of many collagens together with integrin, fibronectin, and vitronectin are altered in the pathway linked to ECM-receptor interactions (80). Thus, our pathways analysis supports the notion that mechanical loading stimulates remodeling of ECM. Note that previous mouse studies using four-point bending have identified signaling pathways linked to EGF receptors, fibronectins, and proteolysis (73), where fibronectins and proteolysis are involved in remodeling of ECM. Third, TGF-β signaling is known to influence diverse processes in embryogenesis, angiogenesis, inflammation, and wound healing. It also plays a major role in the development and maintenance of bone metabolism (81). Lastly, Wnt signaling is one of the central pathways in regulating bone formation (82). Mice with a nonfunctional Lrp5 receptor in this pathway respond poorly to mechanical loading with significant reduction in bone formation compared with wild-type controls (19,83). In osteocytes it is reported that up-regulation of the Wnt pathway together with estrogen receptor, insulin-like growth factor-I, and bone morphogenetic protein pathways are involved in shear-induced mechanotransduction (84). Wnt binds to two distinct receptor complexes: a complex of Frizzled and LRP5/6 and another complex of Frizzled and RORs. The binding of Wnt to the receptors activate two classes of signaling pathways: a β-catenin-mediated canonical pathway and a b-catenin-independent non-canonical pathway (85). In the absence of Wnt signaling, glycogen synthase kinase-3b (GSK-3b) phosphorylates β-catenin in the target cells. Adenomatous polyposis coli (APC) and axin act as scaffold proteins allowing the association of GSK-3b with β-catenin. Phosphorylated β-catenin is degraded through the ubiquitin-proteosome pathway. Wnt1 class ligands such as Wnt1 and Wnt3a activate the canonical pathway through the formation of a complex of Wnt, Frizzled, and LRP5 or LRP6. This complex in turn promotes the phosphorylation of GSK- 3b, which inhibits the kinase activity of GSK-3b. Inactivation of GSK-3b induces the accumulation of β-catenin in the target cells, followed by translocation of accumulated b-catenin into the nucleus. The nuclear β-catenin, together with transcription factors, T-cell factor/lymphoid enhancer factor (TCF/LEF) family members, induces the expression of the Wnt target genes. In the other pathway, Wnt5a binds to a receptor complex of Frizzled and ROR1/2. The binding of Wnt5a to a receptor complex activates heterotrimeric G proteins, which increase intracellular calcium via protein kinase C (PKC) - and calcineurin-dependent mechanisms (86-89). Wnt5a also activates the planar cell

- 115- Chapter 4: Load-induced differential regulation mRNA of trabecular osteocytes polarity pathway through Rho- and Rac/c-Jun amino-terminal kinase (JNK)-dependent signals (26,90-93). In conclusion, we have examined the in vivo effect of mechanical loading on differentially expressed genes in the whole genome derived from trabecular osteocytes, and identified a number of genes and pathways that may play important roles in mediating the skeletal anabolic response to mechanical force. The current study demonstrates that mechanical loading potentially induces multiple signaling pathways involved in mechanotransduction and bone metabolism. Future studies on these unknown genes and signal molecules will provide a better understanding of the molecular pathways involved in mediating the skeleton’s anabolic response to mechanical stress.

4.4. Confirmation of individual load-regulated genes in single loading

Materials and Methods

We used quantitative reverse transcriptase polymerase chain reaction (Real-Time RT-PCR) for selected genes to confirm changes in transcription observed in our differential microarray analysis, using six RNA samples from both control and 8N-loaded groups which were extracted after the acute loading study (described in Chapter 4.1). Real-Time PCR was performed to evaluate mRNA levels for four significantly up-regulated genes observed in microarray data in both single and repeated mechanical loadings, including Wnt5a, Asporin, Xiap and Stat5a. Total RNA (150 pg) was reverse-transcribed into cDNA and amplified by using sensitive for low amount of template RNA SuperScript III Platinum One-Step Quantitative RT-PCR kit with Rox (Invitrogen Inc., Carlsbad, CA). Real-Time PCR was carried out in a 96-well plate using ABI PRISM 7900HT FAST sequence detection system (Applied Biosystems Inc., Foster City, CA). All biological samples were run in duplicates along with primers for housekeeping gene glyceraldehyde-3-phosphate dehydrogese (GAPDH), as a reference gene to normalize the expression data for each gene. Total volume for each reaction was 25 μl. We used TaqMan® primer probes (Assay on Demand, Applied Biosystems Inc.) for GAPDH (assay ID: Mm99999915_g1), Wnt5a (Mm00437347_m1), Asporin (Mm00445945_m1) Xiap (Mm00776505_m1) and Stat5a (Mm00839861_m1) genes in concentration of 10 μM each. Primers were designed for each gene that primed in separate exons and spanned at least one intron to avoid contaminating amplification from genomic DNA. The thermal cycling conditions for Real-Time PCR were: 15 min hold at 50°C (cDNA synthesis), 2 min hold at 95°C, followed by 45 cycles of 95°C for 15 seconds, and 60°C for 30 seconds. Real-Time PCR validation was carried out using the 2-ΔΔCT method (Livak KJ 2001). Normalized gene

- 116- Chapter 4: Load-induced differential regulation mRNA of trabecular osteocytes expression values for each gene based on cycle threshold (CT) values for each of the genes and housekeeping gene GAPDH were generated. Mean ± standard error (SE) values were generated from six samples from each group of either the loaded or control samples tested. The statistical significance of the differences between the means of loaded and control group gene expression values was determined using two-tailed Student’s t-test. The critical value for significance was chosen as P < 0.05.

Results

The mRNA levels of four genes (Wnt5a, Asporin, Xiap and Stat5a), whose up-regulation in both single and repeated loading studies was identified with microarrays, were evaluated by quantitative real-time PCR using six pairs of loaded and non-loaded samples of trabecular osteocytes. The results were consistent with the microarray data. Results of mRNA expression for two genes, including Wnt5a and Asporin (3.14 and 2.41 fold-change respectively), reached statistical significance (P < 0.05) for differential expression level between loaded and control samples. No significant change was found for Xiap and Stat5a mRNA expression levels, however, with a trend to up-regulation (Fig. 4.7).

Real-time PCR

0,014

T * 0,012

0,01

0,008 non-loaded 0,006 * loaded

0,004

0,002 Relative mRNA expression, ddC expression, mRNA Relative

0 Asporin Wnt5a Xiap Stat5a

Fig. 4.7: Quantitative Real-Time PCR analysis for Asporin, Wnt5a, Xiap and Stat5a mRNA transcripts following single mechanical loading.

Real-Time PCR results indicated that the mRNA levels of Wnt5a and Asporin were indeed confirmed to be up-regulated, based on microarray data, between loaded and control specimens in

- 117- Chapter 4: Load-induced differential regulation mRNA of trabecular osteocytes single loading samples. Further confirmation by real-time PCR analysis of individual genes after repeated loading should provide additional information on events that take place in a sustained loading regime.

Discussion

In this study we observed small gene expression ratios in real-time PCR analyses of load-regulated genes in microarrays. These are probably contributed by the complex mix of osteocytic cells being assayed, along with the subtle changes to bone that are observed in response to mechanical loading. Microarray analysis is able to reliably detect small (< 2-fold) changes that prove to be biologically relevant (87), and in our study we were able to confirm two of the differentially expressed genes by real-time PCR analysis, including Wnt5a and asporin. Furthermore, the power of the microarray analysis approach lies in its ability to detect genome-wide, coordinated, or similarly regulated differential gene expression, pointing to perturbed signalling pathways and importantly downstream molecular processes. Our study has identified such relationships between commonly regulated target genes (via WNT signalling pathways) that play roles, in particular, in osteocytes, potentially influencing bone formation, mineralization and remodelling.

- 118- Chapter 4: Load-induced differential regulation mRNA of trabecular osteocytes References 1. Webster D, Morley PL, van Lenthe GH, and Müller R. 2008 A novel in vivo mouse model for mechanically stimulated bone adaptation – a combined experimental and computational validation study. Comput Methods Biomech Biomed Engin 11(5):435-41.

2. Webster D, Wasserman E, Weber F, Bab I, and Müller R. 2010 Mechanical loading of mouse caudal vertebrae increases trabecular and cortical bone mass – dependence on dose and genotype." Biomech Model Mechanobiol., published on-line 30 March

3. Rodan GA 1997 Bone mass homeostasis and biphosphonate action. Bone 20(1):1-4.

4. Evans WJ 1998 Exercise and nutritional needs of elderly people: effects on muscle andbone. Gerodontology 15(1)15-24.

5. Duncun RL, Turner CH 1995 Mechanotranduction and the functional response of bone to mechanical strain. Calcif Tissue Int 57(5):344-358.

6. Biewener AA, Fazzalari NL, Konieczynski DD, Baudinette RV. 1996 Adaptive changes in trabecular architecture in relation to functional strain patterns and disuse. Bone 19(1):1-8.

7. Plas A van der, Aarden EM, Feijen JH, Boer AH de,Wiltink A, Alblas MJ, Leij L de, Nijweide PJ 1994 Characteristics and properties of osteocytes in culture. J Bone Miner Res 9:1697–704.

8. Bonewald LF. 2006 Mechanosensation and transduction in osteocytes. Bonekey Osteovision 3(10):7-15.

9. Burger EH, Klein-Nulend J 1999 Mechanotransduction in bone: role of the lacuno-canalicular network. FASEB J 13 Suppl:S101–S112

10. Bonewald LF. Osteocytes. In: Marcus DFR, Nelson D, Rosen C, editors. Osteoporosis. 3rd Edition. Elsevier; 2007.

11. Bonewlad LF. 2006 Osteocytes as multifunctional cells. J Musculoskelet Neuronal Interact 6(4):331-333.

12. Lanyon LE. 1993 Osteocytes, strain detection, bone modeling and remodeling. Calcif Tissue Int;53 (Suppl 1):S102–6. S106–7.

13. Chambers TJ, Evans M, Gardner TN, Turner-Smith A, Chow JW. 1993 Induction of bone formation in rat tail vertebrae by mechanical loading. Bone Mineral 20(2):167-178.

14. Xing W, Baylink D, Kesavan C, Hu Y, Kapoor S, Chadwick RB, Mohan S 2005 Global geneexpression analysis in the bones reveals involvement of several novel genes and pathways in mediating an anabolic response of mechanical loading in mice. J Cell Biochem 96(5):1049- 60.

15. Lau KH, Kapur S, Kesavan C, Baylink DJ 2006 Up-regulation of the Wnt, estrogen receptor, insulin-like growth factor-I, and bone morphogenetic protein pathways in C57BL/6J osteoblasts as opposed to C3H/HeJ osteoblasts in part contributes to the differential anabolic response to fluid shear. J Biol Chem 281(14):9576-88.

- 119- Chapter 4: Load-induced differential regulation mRNA of trabecular osteocytes 16. Kesavan C, Mohan S, Oberholtzer S, Wergedal JE, Baylink DJ 2005 Mechanical loading- induced gene expression and BMD changes are different in two inbred mouse strains. J Appl Physiol 99(5):1951-7.

17. Frost HM 1987 The mechanostat: a proposed pathogenic mechanism of osteoporoses and the bone mass effects of mechanical and nonmechanical agents. Bone Miner 2(2):73-85.

18. Albright J 1987 The scientific basis of orthopaedics, 2 ed. Appleton-Century Crofts, New York.

19. Robling AG, Niziolek PJ, Baldridge LA, Condon KW, Allen MR, Alam I, Mantila S, Gluhak- Heinrich J, Bellido T, Harris SE and Turner CH. 2007 Mechanical stimulation of bone in vivo reduces osteocyte expression of SOST/Sclerostin. J Biol Chemistry 283(9):5866-5875.

20. Brown P, Botstein D. 1999 Exploring a new world of the genome with DNA microarrays. Nat Genet 21(Suppl. 1):33-37.

21. Eisen MB, Spellman PT, Brown PO, Botstein D. 1998 Cluster analysis and display of genome- wide expression patterns. PNAS 95:14863-14868.

22. Gross TS, Srinivasan S, Liu CC, Clemens TL and Bain SD. 2002 Non-invasive loading of the murine tibia: an in vivo model for the study mechanotransduction. J Bone Miner Res 17(3):493- 501.

23. Rosen CJ, Ackert-Bicknell C, Beamer WG, Nelson T, Adamo M, Cohen P, Bouxsein ML, Horowitz MC. 2004 Allelic differences in a quantitative trait locus affecting insulin-like growth factor-I impact skeletal acquisition and body composition. Pediatr Nephrol 20:255-260.

24. Zhao GH et al. 2000 Targeted overexpression of insulin-like growth factor-I to osteoblasts of transgenic mice: increased trabecular bone volume without increased osteoblast proliferation. Endocrinology 141(7):2674-2682.

25. Lean JM, Mackay AG, Chow JW, Chambers TJ. 1996 Osteocytiv expression of mRNA for c- fos and IGF-1: an immediate early gene response to an osteogenic stimulus. Am J Physiol 270(6 Pt 1):E937-945.

26. Oishi I, Suzuki H, Onishi N, Takada R, Kani S, Ohkawara B. 2003 The receptor tyrosine kinase Ror2 is involved in non-canonical Wnt5a/JNK signalling pathway. Genes Cells 8:645-654.

27. Bonewald LF and Johnson ML. 2008 Osteocytes, mechanosensing and Wnt signaling. Bone 42:606-615.

28. Santos A, Bakker AD, Zandieh-Doulabi B, Semeins CM, Klein-Nulend J. 2009 Pulsating fluid flow modulates gene expression of proteins involved in Wnt signaling pathways in osteocytes. J Orthop Res. Published on-line 7 April 2009.

29. Vaes VLT et al. 2005 Microarray analysis reveals expression regulation of Wnt antagonists in differentiating osteoblasts. Bone 36:803-811.

- 120- Chapter 4: Load-induced differential regulation mRNA of trabecular osteocytes 30. Trebec DP, Chandra D, Gramoun A, Li K, Heersche JN, Manolson MF. 2007 Increased expression of activating factors in large osteoclasts could explain their excessive activity in osteolytic diseases. J Cell Biochem 101(1):205-220.

31. Kalamajski S, Aspberg A, Lindblom K, Heinegard D, Oldberg A. 2009 Aporin competes with decorin for collagen binding, binds calcium, and promotes osteoblast collagen mineralization. Biochemical Journal, published on-line 9 July 2009.

32. Yoshiko Y, Maeda N, and Aubin JE. 2003 Stanniocalcin 1 stimulates osteoblast differentiation in rat calvaria cell cultures. Endocrinology 144(9):4134-4143.

33. Liu Y, Porta A, Peng X, Gengaro K, Cunninghem EB, Li H, Dominguez L, Bellido T, Christakos S. 2004 Prevention of glucocorticoid-induced apoptosis in osteocytes and osteoblasts by Calbindin-D28k. J Bone Miner Res 19(3):479-490.

34. Isaksson OG, Jansson JO, Gause IA 1982 Growth hormone stimulates longitudinal bone growth directly. Science 216:1237–1239.

35. Madsen K, Friberg U, Roos P, Eden S, Isaksson O 1983 Growth hormone stimulates the proliferation of cultured chondrocytes from rabbit ear and rat rib growth cartilage. Nature 304:545–547.

36. Sandstedt J, Tornell J, Norjavaara E, Isaksson OG, Ohlsson C 1996 Elevated levels of growth hormone increase bone mineral content in normal young mice, but not in ovariectomized mice. Endocrinology 137:3368–3374.

37. Clement-Lacroix PA, Ormandy C, Lepescheux L, Ammann P, Damotte D, Goffin V, Bouchard B, Amling M, Gaillard-Kelly M, Binart N, Baron R, Kelly PA 1999 Osteoblasts are a new target for prolactin: analysis of bone formation in prolactin receptor knockout mice. Endocrinology 140:96–105

38. LeBaron M, Ahonen T, Nevalainen M, Rui H. 2007 In vivo response-based identification of direct hormone target cell populations using high-density tissue arrays. Endocrinology 148(3):989-1008.

39. Gong Y, Slee RB, Fukai N, et al. 2001 LDL receptor-related protein 5 (LRP5) affects bone accrual and eye development. Cell 107:513–523.

40. Little RD, Carulli JP, Del Mastro RG, et al. 2002 A mutation in the LDL receptor-related protein 5 gene results in the autosomal dominant high-bone-mass trait. Am J Hum Genet 70:11– 19.

41. Huang H, He X. 2008 Wnt/beta-catenin signaling: new (and old) players and new insights. Curr Opin Cell Biol 20:119–125.

42. Gordon MD, Nusse R. 2006 Wnt signaling: multiple pathways, multiple receptors, and multiple transcription factors. J Biol Chem 281:22429–22433

43. Bennett CN, Longo KA, Wright WS, et al. 2005 Regulation of osteoblastogenesis and bone mass by Wnt10b. Proc Natl Acad Sci USA 102:3324–3329.

- 121- Chapter 4: Load-induced differential regulation mRNA of trabecular osteocytes 44. Andrade A, Nilsson O, Barnes KM, Baron J. 2007 Wnt gene expression in the post-natal growth plate: regulation with chondrocyte differentiation. Bone 40:1361–1369.

45. Rodda SJ, McMahon AP. 2006 Distinct roles for Hedgehog and canonical Wnt signaling in specifi cation, differentiation and maintenance of osteoblast progenitors. Development 133:3231–3244.

46. Dong YF, Soung Y, Schwarz EM, et al. 2006 Wnt induction of chondrocyte hypertrophy through the Runx2 transcription factor. J Cell Physiol 208:77–86.

47. Bennett CN, Ouyang H, Ma YL, et al. 2007 Wnt10b increases postnatal bone formation by enhancing osteoblast differentiation. J Bone Miner Res 22:1924–1932.

48. Glass DA, Bialek P, Ahn JD, et al. 2005 Canonical Wnt signaling in differentiated osteoblasts controls osteoclast differentiation. Dev Cell 8:751–764.

49. Tu X, Joeng K, Nakayama KI, et al. 2007 Noncanonical Wnt signaling through G protein-linked PKCdelta activation promotes bone formation. Dev Cell 12:113–127.

50. Bergenstock MK, Partridge NC. 2007 Parathyroid hormone stimulation of noncanonical Wnt signaling in bone. Ann N Y Acad Sci 1116:354–359.

51. Chang J, Sonoyama W, Wang Z, et al. 2007 Noncanonical Wnt-4 signaling enhances bone regeneration of mesenchymal stem cells in craniofacial defects through activation of p38 MAPK. J Biol Chem 282:30938–30948.

52. Takada I, Suzawa M, Matsumoto K, Kato S. 2007 Suppression of PPAR transactivation switches cell fate of bone marrow stem cells from adipocytes into osteoblasts. Ann N Y Acad Sci 116:182–195.

53. Yamaguchi TP, Bradley A, McMahon AP, Jones S. 1999 A Wnt5a pathway underlies outgrowth of multiple structures in the vertebrate embryo. Development 126:1211–1223.

54. Almeida M, Han L, Bellido T, et al. 2005 Wnt proteins prevent apoptosis of both uncommitted osteoblast progenitors and differentiated osteoblasts by beta-catenin-dependent and - independent signaling cascades involving Src/ERK and phosphatidylinositol 3-kinase/AKT. J Biol Chem 280:41342–41351.

55. Young W et al. 2005 Dentin matrix protein 1 gene cis-regulation: use in osteocytes tochracterize local responses to mechanical loading in vitro and in vivo. J Biol Chem 280(21):296-321.

56. Toyosawa S, Shintani S, Fujivara T, Ooshima T, Sato A, Ijuhin N, and Komori T. 2001 J Bone Miner Res 16:2017-2026.

57. Toyosawa S, Okabayashi K, Komori T, and Ijuhin N. 2004 Bone 34:124-133.

58. Gluhak-Heinrich J, Ye L, Bonewald LF, Feng J, MacDougall M, Harris SE and Pavlin D. 2003 J Bone Miner Res 18:807-817.

59. He G, and George A. 2004 J Biol Chem 279:11649-11656.

- 122- Chapter 4: Load-induced differential regulation mRNA of trabecular osteocytes 60. Amir LR, Jovanovic A, Perdijk F, Toyosawa S, Everts V, and Bronckers A. 2007 Immunolocalization of sibling and RUNX2 proteins during vertical distraction osteogenesis in the human mandible. J Histochem Cytochem 55:1095-1104.

61. Ziros P, Basdra E, and Papavassiliou A. 2008 RUNX2: of bone and stretch.IJBCB 40:1659- 1663.

62. Blanchard A, Ohanian V, Critchley D. 1989 The structure and function of alpha-actinin. J Muscle Res Cell Motil 10:280-289.

63. Jaasma M. 2004 Osteoblast mechanical behaviour and its adaptaion to mechanical loading. Ph.D., Berkely, UC Berkely.

64. Knothe Tate ML, Knothe U, Niederer P. 1998 Experimental elucidation of mechanical load- induced fluid-flow and its potential role in bone metabolism and functional adaptation. Am J Med Sci 316:189-195.

65. Jackson WM, Jaasma MJ, Baik AD and Keaveny TM. 2008 Over-expression of alpha-actinin with a GFP fusion protein is sufficient to increase whole-cell stiffness in human osteoblasts. Ann Biomed Eng 36(10):1605-1614.

66. Tusher VG, Tibshirani R, Chu G. 2001 Significance analysis of microarrays applied to the ionizing radiation response. Proc Natl Acad Sci USA 98:5116-5121.

67. Altman RB, Raychaudhuri SA. 2001 Whole-genome expression analysis: challenges beyond clustering. Curr Opin Struct Biol 11:340-347.

68. Ekins S, Bugrim A, Nikolsky Y, and Nikolskaya T. 2005 Systems biology: applications in drug discovery. In Drug Discovery Handbook (Gad, S. C., ed.), Wiley, New York, pp. 123–183.

69. Nikolsky Y, Nikolskaya T, and Bugrim A. 2005 Biological networks and analysis of experimental data in drug discovery. Drug Discov Today 10:653–662.

70. Ekins S, Nikolsky Y, Bugrim A, Kirillov E, and Nikolskaya T. 2007 Pathway mapping tools for analysis of gigh content data. Methods Mol Biol 356:319-350.

71. Danciu TE, Adam R, Naruse K, Freeman MR, Hauschka PV. 2003 Calcium regulates the PI3K- Akt pathway in stretched osteoblasts. FEBS Lett 536:193–197.

72. Pines A, Romanello M, Cesaratto L, Damante G, Moro L, D’Andrea P, Tell G. 2003 Extracellular ATP stimulates the early growth response protein 1 (Egr-1) via a protein kinase C- dependent pathway in the human osteoblastic HOBIT cell line. Biochem J 373:815–824.

73. Hing W, Baylink D, Kesavan C, Hu Y, Kapoor S, Chadwik R, and Mohan S. 2005 Global gene expression analysis in the bone reveals involvement of several novel genes and pathways in mediating an anabolic response of mechanical loading in mice. J Cell Biochem 96:1049-1060.

74. Zhang P, Turner CH, Yokota H. 2009 Joint loading-driven bone formation and signaling pathways predicted from genome-wide expression profiles. Bone 44:989-998.

- 123- Chapter 4: Load-induced differential regulation mRNA of trabecular osteocytes 75. Harrington LS, Findlay GM, Lamb RF. 2005 Restraining PI3K: mTOR signalling goes back to the membrane. Trends Biochem Sci 30:35–42.

76. Fowlkes JL, Thrailkill KM, Liu L, Wahl EC, Bunn RC, Cockrell GE, et al. 2006 Effects of systemic and local administration of recombinant human IGF-I (rhIGF-I) on de novo bone formation in an aged mouse model. J Bone Miner Res 21:1359–1366.

77. Wang YI, Nishida SA, Elalieh HZ, Long RK, Halloran BP, Bikle DD. 2006 Role of IGF-I signaling in regulating osteoclastogenesis. J Bone Miner Res 21:1350–1358.

78. Leach RM, Richards MP, Praul CA, Ford BCT, McMurtry JP. 2007 Investigation of the insulin- like growth factor system in the avian epiphyseal growth plate. Domest Anim Endocrinol 33:143-53.

79. Cantley LC. 2002 The phosphoinositide 3-kinase pathway. Science 296:1655-7.

80. Lallier TE, Yukna R, Moses RL. 2001 Extracellular matrix molecules improve periodontal ligament cell adhesion to anorganic bone matrix. J Dent Res 80:1748-52.

81. Janssens K, ten Dijke P, Janssens S, Van Hul W. 2005 Transforming growth factor-beta1 to the bone. Endocr Rev 26:743-74.

82. Baron R, Rawadi G. 2007 Targeting the Wnt/beta-catenin pathway to regulate bone formation in the adult skeleton. Endocrinology 148:2635-43.

83. Sawakami K, Robling AG, Ai M, Pitner ND, Liu D, Warden SJ, et al. 2006 The Wnt coreceptor LRP5 is essential for skeletal mechanotransduction but not for the anabolic bone response to parathyroid hormone treatment. J Biol Chem 281:23698-711.

84. Lau KH, Kapur S, Kesavan C, Baylink DJ. 2006 Up-regulation of the Wnt, estrogen receptor, insulin-like growth factor-I, and bone morphogenetic protein pathways in C57BL/6J osteoblasts as opposed to C3H/HeJ osteoblasts in part contributes to the differential anabolic response to fluid shear. J Biol Chem 281:9576-88.

85. Gordon MD, Nusse R. 2006 Wnt signaling: multiple pathways, multiple receptors, and multiple transcription factors. J Biol Chem 281:22429-33.

86. Weeraratna AT, Jiang Y, Hostetter GL, Rosenblatt K, Duray P, Bittner M, et al. 2002 Wnt5a signaling directly affects cell motility and invasion of metastatic melanoma. Cancer Cell 1:279- 288.

87. Dejmek J, Safholm A, Kamp Nielsen C, Andersson T, Leandersson K. 2006 Wnt-5a/Ca2+- induced NFAT activity is counteracted by Wnt-5a/Yes-Cdc42-casein kinase 1alpha signaling in human mammary epithelial cells. Mol Cell Biol 26:6024-6036.

88. Ishitani T, Kishida S, Hyodo-Miura J, Ueno N, Yasuda J, Waterman M, et al. 2003 The TAK1- NLK mitogen-activated protein kinase cascade functions in the Wnt-5a/Ca(2+) pathway to antagonize Wnt/beta-catenin signaling. Mol Cell Biol 23:131-139.

- 124- Chapter 4: Load-induced differential regulation mRNA of trabecular osteocytes 89. Sheldahl LC, Slusarski DC, Pandur P, Miller JR, Kühl M, Moon RT. 2003 Dishevelled activates Ca2+ flux, PKC, and CamKII in vertebrate embryos. J Cell Biol 161:769-777.

90. Qian D, Jones C, Rzadzinska A, Mark S, Zhang X, Steel KP, et al. 2007 Wnt5a functions in planar cell polarity regulation in mice. Dev Biol 306:121-133.

91. Boutros M, Paricio N, Strutt DI, Mlodzik M. 1998 Dishevelled activates JNK and discriminates between JNK pathways in planar polarity and wingless signaling. Cell 94:109-118.

92. Li L, Yuan H, Xie W, Mao J, Caruso AM, Sussman DJ, et al. 1999 Dishevelled proteins lead to two signaling pathways. Regulation of LEF-1 and c-Jun N-terminal kinase in mammalian cells. J Biol Chem 274:129-134.

93. Kobayashi Y, Maeda K, Takahashi N. 2008 Roles of Wnt signaling in bone formation and resorption. Japanese Dental Science Review 44:76-82.

94. Cormen TH, Leiserson CE, Rivest RL, Stein C. 2001 Section 24.3: “Dijkstra’s algorithm”. Introduction to algorithms (Second ed.):pp. 595-601.

- 125- Chapter 5: Synthesis Chapter 5

- 126- Chapter 5: Synthesis Synthesis

5.1. Background and innovations Osteoporosis is a disease characterized by an excessive decrease in bone mass leading to an increased susceptibility to skeletal fracture and deformation, symptoms which can have a dramatic, negative impact on the quality of a person’s life and which, in more extreme cases, can lead to death. A common misconception about this disease is that it is considered to only afflict females but the prevalence in men also increases exponentially with age. The rise in hip fracture rate occurs about 10 years earlier in women than men. By the age of 90, about 17% of males have had a hip fracture, compared to 32% of females. Further to the obvious costs on health, osteoporosis is a global problem and carries with it significant socio-economic costs. This is illustrated by the IOF audit report “Call to Action” published in 2001, which claims that osteoporosis costs national treasuries in the EU over 4.8 billion Euro annually in hospital healthcare alone. Clinically approved strategies aimed at treating the disease employ hormonal based medications which disrupt the bone remodeling process via provocation of bone forming cells or the inhibition of bone resorbing cells. However, these strategies have limited effects and in some cases negative consequences. Medical research is now attempting to target the genes which define osteoporosis using the mouse as a model system for human diseases. Owing to the recent deciphering of the mouse genome and the high homology that exists between the human and mouse genomes, inbred strains of mice represent ideal models for genetic studies. Using the mouse to identify genes implicated in the bone remodeling process could therefore lead to advances in understanding that enable the precise regulation of the genes and proteins responsible for particular bone phenotypes i.e. bone mineral density or bone strength. One interesting phenotype under investigation is the response of bone to mechanical loading or its ‘mechano-sensitivity’. Mechanical loading is the most important physiological/environmental factor regulating bone mass and shape. It has been demonstrated in humans that cyclic overloading enhances bone mass in both cortical and trabecular components. An understanding of the biological pathways (from gene expression to protein function) governing load stimulated bone formation could therefore provide opportunities to mimic or augment bone mechano-sensitivity using pharmacological agents thereby leading to the development of novel strategies in the management of osteoporosis. The goal of this thesis was to propose a two-step strategy which calls for demonstrating (i) a working technique for the isolation of intact total RNA from a well-defined trabecular osteocyte enriched cell population derived from a single mouse caudal vertebra by sequential enzymatic

- 127- Chapter 5: Synthesis digestion with subsequent bone tissue pulverization; (ii) single and repeated load-induced differentially expressed genes in trabecular osteocytes in order to elucidate the molecular mechanisms and pathways involved in the trabecular adaptation to mechanical loads. This novel work proposes a new experimental method of mechanically loaded trabecular bone which combining the isolation of trabecular osteocytic RNA to study load-regulated differential gene expression globally using cDNA microarrays, and use of the single mouse for one biological sample, the only mammalian species with a well-defined genome currently accessible to routine manipulations. A successful implementation of this strategy provides a method and baseline for further studies elucidating the molecular mechanisms involved in the trabecular adaptation to mechanical loads. This project also identified several genes and signaling pathways of interest as targets for future exploration by genetic manipulations, such as knockout, mutated dominant negative and over- expression transgenic technologies for uncovering molecular mechanisms that translate mechanical loads into improved cancellous bone properties. The actual design and implementation of such specific transgenic approaches were beyond the scope of this thesis due to its definition as a feasibility project.

5.2. Developing a method for isolation osteocyte RNA

Chapter 3 describes a new method for the extraction of a sufficient amount of intact total RNA of well-defined trabecular osteocytic population derived from a single murine caudal vertebra (C5). This was perhaps the most ambitious part of the project, aimed at establishing a protocol for the isolation of representative samples of mouse vertebral RNA derived selectively from trabecular osteoblasts/lining cells and osteocytes. The first step in this process was to physically excavate the trabecular bone from the medullar cavity of the target caudal vertebra without contamination from cortical bone. This first step was the most important in determining the yield of the final amount of mRNA. The greater the amount of extracted RNA the better the signal achieved for gene expression. The method and tools employed therefore enabled the maximum extraction of trabecular bone. The medullar cancellous bone (total of approximately 8 mm3) was mechanically separated using a sterile needle with syringe and MicroDrill, and flushed into cold RNAlater for RNA stability. To capture the in vivo expression profile, further ex vivo gene transcription was blocked by adding a transcriptional inhibitor (e.g., actinomycin D). Once the trabecular bone was extracted, the cell populations enriched with bone marrow and osteoblast/lining cells were isolated

- 128- Chapter 5: Synthesis using sequential collagenase digestions and constituted Initial Digest, OBL1 and OBL2 cell fractions. The remaining bony fragments were then briefly grinded in a mortar under liquid nitrogen on dry ice until total pulverization in order to isolate a well-defined population of osteocytes (OST fraction). High quality total RNA preparations were isolated immediately following cell separation using conventional reagents and protocols (TRI Reagent). The accuracy of enzymatic ‘stripping’ of the appropriate cells was confirmed histologically following each digestion step and indicated selective isolation of RNA from osteoblast/lining cell and osteocyte population. Extracted total RNA was then analyzed on quality and quantity using Bioanalyzer technology which has shown intact mRNA suitable for downstream RT-PCR applications. Gene expression profiling and immunohistochmistry analysis were performed in order to confirm appropriate isolation of different cell populations using specific primers for osteoblast and osteocyte lineages. A minimum of o.5 ng of total intact RNA was required for a single cDNA microarray run, using a pico RNA amplification kit. Before attempting to identify load induced gene expression, this protocol was evaluated and optimized by applying it to groups of non-loaded mice. Additionally, in this chapter preliminary trials are described including sequential collagenase digestions with decalcification/resorption agents and a laser-captured microdissection technique which enabled a new approach to be found for developing a robust method for the isolation of representative samples of total RNA derived from well-defined trabecular osteocytes of single mouse caudal vertebra.

5.3. Single and repetitive mechanical loading

Chapter 4 presents the effect of a single and repetitive mechanical loading on differential gene expression and signaling pathway analysis in trabecular osteocytes using cDNA microarrays. Once the protocol for the isolation of total RNA from well-defined trabecular bone cells was shown to yield a sufficient quality and quantity, a critical proof of principle for testing this C57BL/6 mouse model was the identification of load-regulated genes in a trabecular osteocyte-enriched cell population immediately following load application. It was rationalized that a gross gain in bone density as a consequence of multiple load dosing represents the cumulative effect of repetitive single doses. Therefore, to investigate the basic cellular responses to mechanical loading/overloading we studied the set of molecular events evoked by a single loading cycle. The

- 129- Chapter 5: Synthesis intention was to load groups of mice at a magnitude of 0N and 8N for 3’000 cycles, at a frequency of 10 Hz followed by RNA extraction after 6 hours of loading. The microarrays analysis of a single loading revealed that 28000 and 34038 probes, out of a total 45101 probe sets per chip, showed a signal in the 0N and 8N group, respectively. A total of 331 genes whose expression levels showed a 2-fold change between the 0N and 8N groups were identified when the statistical test was carried out at a p <=0.05 level of significance. Of these genes the expression of 281 was up-regulated and that of 50 was down-regulated. In particular, it showed up-regulation of IGF-1 (2.2 fold change), Wnt5a (3.4 fold) and Asporin (3.24 fold) genes which are thought to be activators of osteoblast differentiation and are also responsive to mechanical strain. In contrast, a down-regulation of WNT inhibitor factor 1 gene (WIF-1, 1.8 fold), inhibitor of WNT/beta-cathenin pathway, thought to play an important role in stimulation of osteoblast differentiation and bone formation, was identified. In addition, quantitative real-time PCR results have shown that the mRNA levels of Wnt5a and Asporin were indeed confirmed to be up-regulated, based on microarray data, between loaded and control samples in the acute loading study. The single loading study investigated the basic cellular responses to mechanical loading evoked by a single loading cycle; however the question remained whether these were acute or chronic responses. Therefore, the mechanical loading protocol used in single loading was employed for an extended period of time, 3 times a week for 4 weeks to directly compare results to investigate the response of trabecular osteocytes in single and repeated loading reqimes. The microarray analysis of the repeated loading revealed that 27997 and 36414 probes out of a total of 45101 probe sets per chip, showed a signal in the 0N and 8N group, respectively. A total of 1342 genes whose expression levels showed a 2-fold change between the 0N and 8N groups were identified when the statistical test was carried out at a P <=0.05 level of significance. Of these genes the expression of 781 was up-regulated and that of 561 was down-regulated. In particular, it showed up-regulation of Wnt5a (2.19 fold), Asporin (4.9 fold) and DMP-1 (2.17 fold) genes which are thought to be activators of osteoblast differentiation and are also responsive to mechanical loading. The genes presented after a single load dose, but no longer expressed after the 4-week loading protocol, were considered transient, for example IGF-1. The identification of these genes raised interesting questions on the time scale of their activation throughout the duration of the repeated loading protocol, a consideration for future research. Conversely, genes expressed after the 4-week protocol, but not previously identified in the acute study, were considered chronic, for instance DMP-1, and might also pose interesting questions for future research. A third possible result for some genes was that there were no differences in gene expression between the two loading protocols, Wnt5a and asporin

- 130- Chapter 5: Synthesis genes. This finding confirmed the assumption that the anabolic adaptation of the cancellous bone is truly an additive process resulting from three times per week single doses of 5-minute loading regimes. These results will have implications for the development of loading regimes to induce anabolic bone adaptation, and provide further insight into the fundamental genetic processes governing bone adaptation. Also, the results of Real-Time PCR in Chapter 4 indicate that the mRNA levels of Wnt5a and Asporin genes were indeed confirmed to be up-regulated, based on microarray data, between loaded and control specimens in single mechanical loading samples.

5.4. Functional genomics for identification of load-regulated pathways

Additionally, Chapter 4 presents load-regulated signaling pathway analysis following single and repeated mechanical loads using the MetaCore GeneGO software program. Functional genomics examined the in vivo effect of single and repeated mechanical loading on differentially expressed genes in the whole genome derived from trabecular osteocytes, and identified a number of signaling molecules and pathways that may play important roles in mediating the skeletal anabolic response to mechanical force. MetaCore comes with a built-in natural language processing module MedScan and a comprehensive database containing more than 150,000 events of regulation, interaction and modification between proteins and cell processes obtained from PubMed which allows it to generate a biological association network (BAN) of known protein–protein interactions. By importing microarray expression data into the BAN, co-expressed genes associated with specific signaling pathways were identified. Functional genomics analysis revealed overall 65 load-regulated signaling pathways in a single mechanical loading and 153 load-regulated signaling pathways in a repeated loading study between loaded and control samples. The current study demonstrated that mechanical loading potentially induces multiple signaling pathways involved in mechanotransductuion and bone metabolism. Future studies on these unknown genes and signal molecules will provide a better understanding of the molecular pathways involved in mediating the skeleton’s anabolic response to mechanical stress.

5.5. Limitations

In this thesis, a number of limitations can be identified. The whole procedure for isolation of osteocyte RNA was time-consuming and did not allow RNA extraction from more than 12 mice per day. Mechanical separation of trabecular bone from caudal vertebra did not always yield a sufficient

- 131- Chapter 5: Synthesis amount of 8 mm3 tissue volume to obtain the final amount of extracted 0.5 ng of total RNA. Additionally, isolated osteocytic RNA may have some contamination from osteoblastic lineage which has been shown by the gene expression profiling, described in Chapter 3. Finally, the actual design and implementation of specific transgenic technologies were obviously beyond the scope of this thesis due to its definition as a feasibility project.

5.6. Future work

Future work will be focused on applying knockout, mutated dominant negative and over-expression transgenic technologies to the study of molecular mechanisms involved in the cancellous bone response to mechanical loading. It is anticipated that several genes of interest which were identified in this project will be targets for future exploitation by genetic manipulation strategies to elucidate molecular mechanisms that translate mechanical loads into improved bone properties. To this end, future work will characterize the trabecular bone response to loading in Wnt5a+/- mutated mice using micro-computered tomography and bone histomorph0metry

5.7. Conclusions

In conclusion, this thesis provided a method for the isolation of mRNA from well-defined trabecular osteocytes, and cDNA microarrays revealed the mechanobiological effect of acute and chronic loading regimes in-vivo on global murine differential gene expression, including analysis of signaling pathways. It investigated load-induced basic cellular response in trabecular bone and identified genes of interest that were regulated by mechanical loading in order to elucidate the molecular mechanisms involved in the osteogenic anabolic effect of mechanical loading. This in turn will lead the way to studying the role of genes in load-stimulated bone formation in corresponding transgenic systems.

- 132- Appendix Appendix

Contents

A1. List of up-regulated genes in single loading…………………………………………………..134

A2. List of down-regulated genes in single loading……………………………………………….143

A3. List of up-regulated genes in repeated loading………………………………………………..146

A4. List of down-regulated genes in repeated loading…………………………………………….168

A5. List of load-regulated signalling pathways in single loading………………………………....184

A6. List of load-regulated signalling pathways in repeated loading……………………………....186

- 133- Appendix Table A1. Up-regulated gene expression in trabecular osteocytes induced by single loading dose

Cell growth and differentiation Gene Entrez Fold Gene Description Symbol Gene ID Change Mpo 17523 Myeloperoxidase 5.25* Rb1 19645 Retinoblastoma 1 4.35 Slfn4 20558 Schlafen 4 4.22 Braf 109880 Braf transforming gene 3.97 Wipi1 52639 WD repeat domain phosphoinositide interacting 1 3.87 Jmjd6 107817 Jumonji domain containing 6 3.81 Wnt5a** 22418 Wingless-related MMTV integration site 5A 3.37 Foxc2 14234 Forkhead box C2 3.28 Aspn** 66695 Asporin 3.24 Xiap** 11798 X-linked inhibitor of apoptosis 3.23 Vps13a 271564 Vacuolar protein sorting 13A 3.17 Glucosaminyl (N-acetyl) transferase 2, I-branching Gcnt2 14538 3.13 enzyme Nat12 70646 N-acetyltransferase 12 3.09 Ankrd11** 77087 Ankyrin repeat domain 11 3.05 Ccr1 12768 Chemokine (C-C motif) receptor 1 2.96 Aldh1a7 26358 Aldehyde dehydrogenase family 1, subfamily A7 2.96 Mdm2 17246 Transformed mouse 3T3 cell double minute 2 2.95 Cxadr 13052 Coxsackievirus and adenovirus receptor 2.86 Tbc1d1 57915 TBC1 domain family, member 1 2.80 Vav1 22324 Vav 1 oncogene 2.76 Tgm2 21817 Transglutaminase 2, C polypeptide 2.62 Itk 16428 IL2-inducible T-cell kinase 2.57 Pex5 19305 Peroxisome biogenesis factor 5 2.49 Plxnc1 54712 Plexin C1 2.47 Slfn2 20556 Schlafen 2 2.47 Plek 56193 Pleckstrin 2.47 Abhd5 67469 Abhydrolase domain containing 5 2.42 Dhx36 72162 DEAH (Asp-Glu-Ala-His) box polypeptide 36 2.41 Stc1 20855 Stanniocalcin 1 2.39 V-maf musculoaponeurotic fibrosarcoma oncogene, Maff 17133 2.37 protein F Stk4 58231 Serine/threonine kinase 4 2.34 Pabpn1 54196 Poly(A) binding protein, nuclear 1 2.36 Adaptor protein, phosphotyrosine interaction, Appl2 216190 2.26 leucine zipper 2 Nf1 18015 Neurofibromatosis 1 2.26 Prlr 19116 Prolactin receptor 2.25

- 134- Appendix Golgi associated PDZ and coiled-coil motif Gopc 94221 2.25 containing Ptpro 19277 Protein tyrosine phosphatase, receptor type, O 2.22 Efnb2 13642 Ephrin B2 2.22 Igf1 16000 Insulin-like growth factor 1 2.22 Ldb1 16825 LIM domain binding 1 2.20 Bin1 30948 Bridging integrator 1 2.18 Hrb 15463 HIV-1 Rev binding protein 2.17 Ddx27 228889 DEAD (Asp-Glu-Ala-Asp) box polypeptide 27 2.16 Kng1 16644 Kininogen 1 2.15 Zfx 22764 Zinc finger protein X-linked 2.14 Mgll 23945 Monoglyceride lipase 2.13 Mef2c** 17260 Myocyte enhancer factor 2C 2.09 Itga4 16401 Integrin alpha 4 2.08 Xdh** 22436 Xanthine dehydrogenase 2.07 Adam10 11487 A disintegrin and metallopeptidase domain 10 2.04 Ifi204** 15951 Interferon activated gene 204 2.03 Hoxb4 15412 Homeo box B4 2.03 Ets2 23872 E26 avian leukemia oncogene 2, 3' domain 2.03 Sprouty protein with EVH-1 domain 1, related Spred1 114715 2.02 sequence Ppp3cb 19056 Protein phosphatase 3 catalytic subunit, β-isoform 2.01 Dhx30 72831 DEAH (Asp-Glu-Ala-His) box polypeptide 30 2.01 *p value smaller than 0.05 **Genes reportedly associated with bone metabolism

Apoptosis Camk1d 227541 Calcium/calmodulin-dependent protein kinase ID 2.99* Luc7l 66978 Luc7 homolog (S. cerevisiae-like) 2.77 Rtn4 68585 Reticulon 4 2.76 GULP, engulfment adaptor PTB domain containing Gulp1 70676 2.60 1 Tmem19 67226 Transmembrane protein 19 2.58 Rcsd1 226594 RCSD domain containing 1 2.55 Pmaip1 58801 Phorbol-12-myristate-13-acetate-induced protein 1 2.50 Tmem167 66074 Transmembrane protein 167 2.34 Zfyve26 211978 Zinc finger, FYVE domain containing 26 2.19 Becn1 56208 Beclin 1, autophagy related 2.18 Cytotoxic granule-associated RNA binding protein Tia1 21841 2.13 1 Ncf1 17969 Neutrophil cytosolic factor 1 2.09 Tumor necrosis factor receptor superfamily, Tnfrsf1b 21938 2.08 member 1b Tmem77 67171 Transmembrane protein 77 2.05

- 135- Appendix

Thyn1 77862 Thymocyte nuclear protein 1 2.01 *p value smaller than 0.05

Cell cycle Btbd11** 74007 BTB (POZ) domain containing 11 3.91* Clasp1 76707 CLIP associating protein 1 3.21 Oligonucleotide/oligosaccharide-binding fold Obfc2a 109019 containing 2A 2.90 Ccnd3 12445 Cyclin D3 2.45 Gas5 14455 Growth arrest specific 5 2.42 Psmd11 69077 Proteasome 26S subunit, non-ATPase, 11 2.30 Rmnd5a 68477 Required for meiotic nuclear division 5 homolog A 2.29 Calm1 12313 Calmodulin 1 2.26 Stag1 20842 Stromal antigen 1 2.26 Rfc3 69263 Replication factor C (activator 1) 3 2.18 Ppp6c 67857 Protein phosphatase 6, catalytic subunit 2.03 Gnl3 30877 Guanine nucleotide binding protein-like 3 2.01 Kpna4 16649 Karyopherin (importin) alpha 4 2.01 *p value smaller than 0.05 **Genes reportedly associated with bone metabolism

Signal transduction Cdgap 12549 Cdc42 GTPase-activating protein 4.00* Adrbk2 320129 Adrenergic receptor kinase, beta 2 3.68 Ras association (RalGDS/AF-6) domain family Rassf5 54354 2.88 member 5 Als2cl 235633 ALS2 C-terminal like 2.87 Ly6g6c 68468 Lymphocyte antigen 6 complex, locus G6C 2.84 colony stimulating factor 2 receptor beta, Csf2rb 12983 2.72 granulocyte-macrophage Pram1 378460 PML-RAR alpha-regulated adaptor molecule 1 2.68 Rap guanine nucleotide exchange factor (GEF)-like Rapgefl1 268480 2.64 1 Akap10 56697 A kinase (PRKA) anchor protein 10 2.60 Serine (or cysteine) peptidase inhibitor, clade A, Serpina3c 16625 2.56 member 3C Stat5a** 20850 Signal transducer and activator of transcription 5A 2.49 Ptgfr 19220 Prostaglandin F receptor 2.48 Pdpk1 18607 3-phosphoinositide dependent protein kinase-1 2.42 Iqgap1 29875 IQ motif containing GTPase activating protein 1 2.34 Plxna2 18845 Plexin A2 2.30 Rab23 19335 RAB23, member RAS oncogene family 2.26 Traf1 22029 Tnf receptor-associated factor 1 2.25

- 136- Appendix

Lrba 80877 LPS-responsive beige-like anchor 2.24 Cd19 12478 CD19 antigen 2.18 Ccr9 12769 Chemokine (C-C motif) receptor 9 2.16 Gpr68 238377 G protein-coupled receptor 68 2.15 Stat1 20846 Signal transducer and activator of transcription 1 2.14 Stat4 20849 Signal transducer and activator of transcription 4 2.14 Tm6sf1 107769 transmembrane 6 superfamily member 1 2.11 Akap12 83397 A kinase (PRKA) anchor protein (gravin) 12 2.10 Map3k3 26406 mitogen-activated protein kinase kinase kinase 3 2.09 Rapgef6 192786 Rap guanine nucleotide exchange factor (GEF) 6 2.09 Prdx6 11758 Peroxiredoxin 6 2.07 MARVEL (membrane-associating) domain Marveld1 277010 2.05 containing 1 Lnx2 140887 Ligand of numb-protein X 2 2.01 Muskelin 1, intracellular mediator containing kelch Mkln1 27418 2.00 motifs *p value smaller than 0.05 **Genes reportedly associated with bone metabolism

Transcription Ppih 66101 Peptidyl prolyl isomerase H 4.05* Rbm39 170791 RNA binding motif protein 39 3.92 Trim26 22670 Tripartite motif-containing 26 3.42 Zfp711 245595 Zinc finger protein 711 3.38 Zfp260 26466 Zinc finger protein 260 3.37 Cdyl2 75796 Chromodomain protein, Y chromosome-like 2 3.30 Mycbp 56309 C-myc binding protein 3.01 Suz12 52615 Suppressor of zeste 12 homolog 2.95 Nuclear factor of kappa light polypeptide gene Nfkbiz 80859 2.90 enhancer, zeta Zfp207 22680 Zinc finger protein 207 2.90 Cpsf6 432508 Cleavage and polyadenylation specific factor 6 2.89 Supt6h 20926 Suppressor of Ty 6 homolog 2.79 Rpo1-3 20018 RNA polymerase 1-3 2.78 Rwdd4a 192174 RWD domain containing 4A 2.76 Papola 18789 Poly (A) polymerase alpha 2.73 Cpsf6 432508 Cleavage and polyadenylation specific factor 6 2.68 Prpf40a 56194 PRP40 pre-mRNA processing factor 40 homolog A 2.67 Zfp248 72720 Zinc finger protein 248 2.67 Trps1 83925 Trichorhinophalangeal syndrome I 2.60 Orc2l 18393 Origin recognition complex, subunit 2-like 2.55 Ebf3 13593 Early B-cell factor 3 2.51 Baz1b 22385 Bromodomain adjacent to zinc finger domain, 1B 2.51

- 137- Appendix

Nlrc3 268857 NLR family, CARD domain containing 3 2.45 Polr1e 64424 Polymerase (RNA) I polypeptide E 2.43 Jarid1b 75605 Jumonji, AT rich interactive domain 1B (Rbp2 like) 2.43 Xbp1 22433 X-box binding protein 1 2.41 Rlf 109263 Rearranged L-myc fusion sequence 2.40 Phf23 78246 PHD finger protein 23 2.35 Ccdc128 73825 Coiled-coil domain containing 128 2.29 Limd1 29806 LIM domains containing 1 2.29 Rp9 55934 Retinitis pigmentosa 9 2.27 Sp4 20688 Trans-acting transcription factor 4 2.26 Noc4l 100608 Nucleolar complex associated 4 homolog 2.26 Carboxy-terminal domain, polypeptide A Ctdp1 67655 2.24 phosphatase, subunit 1 Ccnt2 72949 Cyclin T2 2.22 Zfp711 245595 Zinc finger protein 711 2.21 Jmjd1c 108829 Jumonji domain containing 1C 2.20 Trps1 83925 Trichorhinophalangeal syndrome I 2.16 Rnf168 70238 Ring fnger protein 168 2.16 Tcta 102791 T-cell leukemia translocation altered gene 2.15 Frg1 14300 FSHD region gene 1 2.15 Tox4 268741 TOX high mobility group box family member 4 2.14 Egr2 13654 Early growth response 2 2.14 Zfp503 218820 Zinc finger protein 503 2.13 Mobkl2a 208228 MOB1, Mps one binder kinase activator-like 2A 2.10 Hnrpab 15384 Heterogeneous nuclear ribonucleoprotein A/B 2.08 Fusip1 14105 FUS interacting protein (serine-arginine rich) 1 2.07 Ptrf 19285 Polymerase I and transcript release factor 2.04 Nsun6 74455 NOL1/NOP2/Sun domain family 6 2.03 Sf3b2 319322 Splicing factor 3b, subunit 2 2.02 Creb1 12912 cAMP responsive element binding protein 1 2.02 Bbx 70508 Bobby sox homolog 2.01 Rnf169 108937 Ring finger protein 169 2.00 *p value smaller than 0.05

Immune response Vpreb1 22362 Pre-B lymphocyte gene 1 5.70* Msr1 20288 Macrophage scavenger receptor 1 4.10 Il1b 16176 Interleukin 1 beta 3.53 Ccl12 20293 Chemokine (C-C motif) ligand 12 3.51 Gardner-Rasheed feline sarcoma viral oncogene Fgr 14191 3.22 homolog Ccr2 12772 Chemokine (C-C motif) receptor 2 3.17

- 138- Appendix

Chi3l4 104183 Chitinase 3-like 4 3.13 Pla2g7 27226 Phospholipase A2, group VII 3.08 Il1rn 16181 Interleukin 1 receptor antagonist 3.08 Clec4d 17474 C-type lectin domain family 4, member d 2.90 Ier5 15939 Immediate early response 5 2.83 Ifitm6 213002 Interferon induced transmembrane protein 6 2.46 Fcgr2b 14130 Fc receptor, IgG, low affinity IIb 2.25 Ccr2 12772 Chemokine (C-C motif) receptor 2 2.19 Cd300lb 217304 CD300 antigen like family member B 2.19 Rac2 19354 RAS-related C3 botulinum substrate 2 2.07 *p value smaller than 0.05

Muscle contraction Dtna 13527 Dystrobrevin alpha 4.36* Prrx1 18933 Paired related homeobox 1 2.40 Cugbp2 14007 CUG triplet repeat, RNA binding protein 2 2.07 *p value smaller than 0.05

Cytoskeleton remodeling Disintegrin-like metallopeptidase, thrombospondin Adamts4 240913 4.11* type 1, 4 S100a9 20202 S100 calcium binding protein A9 (calgranulin B) 2.91 Tubb6 67951 Tubulin, beta 6 2.91 Cnp 12799 2',3'-cyclic nucleotide 3' phosphodiesterase 2.86 Evl 14026 Ena-vasodilator stimulated phosphoprotein 2.55 Dnajc7 56354 DnaJ (Hsp40) homolog, subfamily C, member 7 2.45 Rai14 75646 Retinoic acid induced 14 2.44 Fmnl1 57778 Formin-like 1 2.16 Tsga14 83922 Testis specific gene A14 2.15 Elmo1 140580 Engulfment and cell motility 1, ced-12 homolog 2.12 Ttll3 101100 Tubulin tyrosine ligase-like family, member 3 2.10 Wdr1 22388 WD repeat domain 1 2.07 Arhgap10 78514 Rho GTPase activating protein 10 2.07 *p value smaller than 0.05

Proteolysis Disintegrin-like metallopeptidase, thrombospondin Adamts1 11504 3.63* type 1, 1 Ermp1 226090 Endoplasmic reticulum metallopeptidase 1 2.39 Asb3 65257 Ankyrin repeat and SOCS box-containing protein 3 2.23 Serine (or cysteine) peptidase inhibitor, clade B, Serpinb1a 66222 2.21 member 1a

- 139- Appendix

Rnf13 24017 Ring finger protein 13 2.19 Lnpep 240028 Leucyl/cystinyl aminopeptidase 2.07 Ubfd1 28018 Ubiquitin family domain containing 1 2.03 *p value smaller than 0.05

Translation Tnrc6a 233833 Trinucleotide repeat containing 6a 3.96* Etf1 225363 Eukaryotic translation termination factor 1 2.45 Secisbp2 75420 SECIS binding protein 2 2.44 Srp54b 665155 Signal recognition particle 54b 2.34 Protein phosphatase 1, regulatory (inhibitor) Ppp1r15b 108954 2.19 subunit 15b Syk 20963 Spleen tyrosine kinase 2.17 Stom 13830 Stomatin 2.13 Rpl37a 19981 Ribosomal protein L37a 2.05 *p value smaller than 0.05

G-protein signaling Dock8 76088 Dedicator of cytokinesis 8 3.97* Fpr-rs2 14289 Formyl peptide receptor, related sequence 2 2.88 Diap1 13367 Diaphanous homolog 1 2.62 Gpr177 68151 G protein-coupled receptor 177 2.33 Gpr137b 83924 G protein-coupled receptor 137B 2.31 Gtpbp4 69237 GTP binding protein 4 2.14 *p value smaller than 0.05

Cell adhesion Sele 20339 Selectin, endothelial cell 3.74* Fermt3 108101 Fermitin family homolog 3 3.19 Cd38 12494 CD38 antigen 2.93 Itsn1 16443 Intersectin 1 (SH3 domain protein 1A) 2.83 Tmem206 66950 Transmembrane protein 206 2.69 Olfm4 380924 Olfactomedin 4 2.45 LIM domain containing preferred translocation Lpp 210126 2.33 partner in lipoma Pleckstrin homology, Sec7 and coiled-coil Pscdbp 227929 2.33 domains, binding protein Killer cell lectin-like receptor, subfamily A, Klra2 16633 2.27 member 2 Pcdhb16 93887 Protocadherin beta 16 2.23 Emilin2 246707 Elastin microfibril interfacer 2 2.10 *p value smaller than 0.05

- 140- Appendix

Chemotaxis Chi3l3 12655 Chitinase 3-like 3 2.93* C5ar1 12273 Complement component 5a receptor 1 2.31 *p value smaller than 0.05

Transport Appbp2 66884 Amyloid beta precursor protei binding protein 2 3.63* Clca1 12722 Chloride channel calcium activated 1 3.48 Solute carrier family 8 (sodium/calcium exchange), Slc8a1 20541 3.11 member 1 Srr 27364 Serine racemase 2.96 Atp8b4 241633 ATPase, class I, type 8B, member 4 2.73 Vps33b 233405 Vacuolar protein sorting 33B 2.67 Xpo4 57258 Exportin 4 2.66 Cep350 74081 Centrosomal protein 350 2.57 Exoc4 20336 Exocyst complex component 4 2.50 Nup188 227699 Nucleoporin 188 2.46 Signal recognition particle receptor ('docking Srpr 67398 2.44 protein') Collagen, type IV, alpha 3 (Goodpasture antigen) Col4a3bp 68018 2.41 binding protein Solute carrier family 25 (mitochondrial carrier), Slc25a25 227731 2.37 member 25 Snx9 66616 Sorting nexin 9 2.29 March1 72925 Membrane-associated ring finger (C3HC4) 1 2.28 Xpo1 103573 Exportin 1, CRM1 homolog (yeast) 2.27 Translocating chain-associating membrane protein Tram2 170829 2.25 2 Solute carrier family 7 (cationic amino acid Slc7a2 11988 2.24 transporter), member 2 Kpna3 16648 Karyopherin (importin) alpha 3 2.19 Ubiquinol-cytochrome c reductase, complex III Uqcrq 22272 2.13 subunit VII Pacs1 107975 Phosphofurin acidic cluster sorting protein 1 2.08 Osbpl7 71240 Oxysterol binding protein-like 7 2.08 Osbpl6 99031 Oxysterol binding protein-like 6 2.04 Ppm1m 67905 Protein phosphatase 1M 2.04 Pleckstrin homology domain-containing, family A Plekha3 83435 2.02 member 3 Ipo9 226432 Importin 9 2.01 *p value smaller than 0.05

Energy metabolism Hdc 15186 Histidine decarboxylase 3.34*

- 141- Appendix

Phca 66190 Phytoceramidase, alkaline 2.92 Pmm2 54128 Phosphomannomutase 2 2.90 Phosphatidylinositol-5-phosphate 4-kinase, type II, Pip4k2a 18718 2.76 alpha Protein kinase, AMP-activated, alpha 1 catalytic Prkaa1 105787 2.73 subunit Lipg 16891 Lipase, endothelial 2.71 Ggps1 14593 Geranylgeranyl diphosphate synthase 1 2.66 Cept1 99712 Choline/ethanolaminephosphotransferase 1 2.43 Ipp 16351 IAP promoted placental gene 2.38 Glb1l 74577 Galactosidase, beta 1-like 2.38 Adc 242669 Arginine decarboxylase 2.38 Angptl4 57875 Angiopoietin-like 4 2.29 Dhodh 56749 Dihydroorotate dehydrogenase 2.26 UDP glucuronosyltransferase 1 family, polypeptide Ugt1a6a 94284 2.19 A6A Glipr2 384009 GLI pathogenesis-related 2 2.19 Tankyrase, TRF1-interacting ankyrin-related ADP- Tnks 21951 2.15 ribose polymerase Agxt2l2 72947 Alanine-glyoxylate aminotransferase 2-like 2 2.14 Atp11c 320940 ATPase, class VI, type 11C 2.06 Gda 14544 Guanine deaminase 2.03 Cytochrome P450, family 2, subfamily d, Cyp2d22 56448 2.02 polypeptide 22 *p value smaller than 0.05

Neurophysiological process Gphn 268566 Gephyrin 2.52* *p value smaller than 0.05

- 142- Appendix Table A2. Down-regulated gene expression in trabecular osteocytes induced by single loading dose

Cell growth and differentiation Gene Entrez Gene Description Fold Change Symbol Gene ID Arhgap24 231532 Rho GTPase activating protein 24 3.32* Gprin3 243385 GPRIN family member 3 3.08 Foxq1 15220 Forkhead box Q1 2.46 Olfm3 229759 Olfactomedin 3 2.35 Foxp2 114142 Forkhead box P2 2.30 Triobp 110253 TRIO and F-actin binding protein 2.22 Sparc/osteonectin, cwcv and kazal-like domains Spock2 94214 2.20 proteoglycan 2 Prdm1 12142 PR domain containing 1, with ZNF domain 2.19 Zfp39 22698 Zinc finger protein 39 2.17 G protein regulated inducer of neurite outgrowth Gprin2 432839 2.09 2 Fzd5 14367 Frizzled homolog 5 2.05 Six4 20474 Sine oculis-related homeobox 4 homolog 2.03 Col19a1 12823 Collagen, type XIX, alpha 1 2.01 *p value smaller than 0.05

Apoptosis Otud1 71198 OTU domain containing 1 2.30* *p value smaller than 0.05

Cell cycle Ubiquitin-like, containing PHD and RING Uhrf2 109113 2.32* finger domains 2 Cct4 12464 Chaperonin subunit 4 (delta) 2.10 Brca2 12190 Breast cancer 2 2.06 Phgdh 236539 3-phosphoglycerate dehydrogenase 2.01 *p value smaller than 0.05

Signal transduction Lrrc28 67867 Leucine rich repeat containing 28 3.16* ATPase, Ca++ transporting, cardiac muscle, Atp2a2 11938 2.49 slow twitch 2 Protein phosphatase 2, regulatory subunit B Ppp2r5a 226849 2.12 (B56), α-isoform Akt2 11652 Thymoma viral proto-oncogene 2 2.02 *p value smaller than 0.05

Transcription

- 143- Appendix

Rian 75745 RNA imprinted and accumulated in nucleus 2.90* Foxf1a 15227 Forkhead box F1a 2.76 Zfp449 78619 Zinc finger protein 449 2.65 Host cell factor C1 regulator 1 (XPO1- Hcfc1r1 353502 2.43 dependent) Hdac11 232232 Histone deacetylase 11 2.32 Smyd3 69726 SET and MYND domain containing 3 2.20 Zdhhc2 70546 Zinc finger, DHHC domain containing 2 2.09 *p value smaller than 0.05

Immune response Zcchc11 230594 Zinc finger, CCHC domain containing 11 2.00* *p value smaller than 0.05

Muscle contraction Ryr2 20191 Ryanodine receptor 2, cardiac 2.04* *p value smaller than 0.05

Cytoskeleton remodeling Mid1 17318 Midline 1 2.91* Ablim3 319713 Actin binding LIM protein family, member 3 2.34 Clns1a 12729 Chloride channel, nucleotide-sensitive, 1A 2.01 *p value smaller than 0.05

Proteolysis Usp25 30940 Ubiquitin specific peptidase 25 2.29* *p value smaller than 0.05

Cell adhesion Cd34 12490 CD34 antigen 2.03* *p value smaller than 0.05

Transport Ndufc1 66377 NADH dehydrogenase (ubiquinone) 1, 1 2.51* Sh3gl2 20404 SH3-domain GRB2-like 2 2.39 Transient receptor potential cation channel, Trpm1 17364 2.36 subfamily M, 1 Exoc8 102058 Exocyst complex component 8 2.32 Filip1 70598 Filamin A interacting protein 1 2.23 *p value smaller than 0.05

Energy metabolism Osbpl6 99031 Oxysterol binding protein-like 6 2.41*

- 144- Appendix

Ppp1r3c 53412 Protein phosphatase 1, regulatory inhibitor, 3C 2.38 Ckm 12715 Creatine kinase, muscle 2.11 Tkt 21881 Transketolase 2.07 *p value smaller than 0.05

Neurophysiological process Txlnb 378431 Taxilin beta 2.50* Asrgl1 66514 Asparaginase like 1 2.10 *p value smaller than 0.05

Unknown process DNA segment, Chr 1, ERATO Doi 399, D1Ertd399e 52296 4.08* expressed C79741 97877 Expressed sequence C79741 3.70 C77717 97361 Expressed sequence C77717 2.74 *p value smaller than 0.05

- 145- Appendix Table A3. Up-regulated gene expression in trabecular osteocytes induced by multiple loading doses

Cell growth and differentiation Gene Entrez Fold Gene Description Symbol Gene ID Change Cryab 12955 Crystallin, alpha B 8.22* H19 14955 H19 fetal liver mRNA 8.03 Tnmd 64103 Tenomodulin 7.75 Ptn** 19242 Pleiotrophin 7.03 Ttc8 76260 Tetratricopeptide repeat domain 8 6.67 Dixdc1 330938 DIX domain containing 1 6.59 Ldb3 24131 LIM domain binding 3 5.78 Aspn** 66695 Asporin 4.91 Thbs4 21828 Thrombospondin 4 4.81 Col3a1 12825 Collagen, type III, alpha 1 4.80 Ttn 22138 Titin 4.76 Gas1 14451 Growth arrest specific 1 4.26 Eno3 13808 Enolase 3, beta muscle 4.24 Mpz 17528 Myelin protein zero 4.15 Tnnt3 21957 Troponin T3, skeletal, fast 4.02 Meox2 17286 Mesenchyme homeobox 2 4.01 Fkbp4 14228 FK506 binding protein 4 4.00 Pleckstrin homology domain, family B (evectins) Plekhb1 27276 3.97 member 1 Cilp** 214425 Cartilage intermediate layer protein 3.80 Itm2a** 16431 Integral membrane protein 2A 3.62 Sema3c 20348 Semaphorin 3C 3.51 Pcp4 18546 Purkinje cell protein 4 3.31 Col9a3 12841 Collagen, type IX, alpha 3 3.30 Rpgr 19893 Retinitis pigmentosa GTPase regulator 3.26 Atm 11920 Ataxia telangiectasia mutated homolog 3.18 Sema3d 108151 Semaphorin 3D 3.16 Itgb1bp2 26549 Integrin beta 1 binding protein 2 3.15 Fmod 14264 Fibromodulin 3.15 Tmem46 219134 Transmembrane protein 46 3.10 Ankrd6 140577 Ankyrin repeat domain 6 3.09 Prkcdbp 109042 Protein kinase C, delta binding protein 3.09 Mapk12 29857 Mitogen-activated protein kinase 12 3.03 Cgrrf1 68755 Cell growth regulator with ring finger domain 1 3.01 Cdh13 12554 Cadherin 13 3.00 Fzd7 14369 Frizzled homolog 7 2.95 Drg1 13494 Developmentally regulated GTP binding protein 1 2.93

- 146- Appendix

Kif2a 16563 Kinesin family member 2A 2.90 Tmem58 77552 Transmembrane protein 58 2.87 Lix1l** 280411 Lix1-like 2.79 Bbs2 67378 Bardet-Biedl syndrome 2 homolog 2.78 Ctnna1 12385 Catenin (cadherin associated protein), alpha 1 2.75 Meg3** 17263 Maternally expressed 3 2.69 Frzb 20378 Frizzled-related protein 2.66 N-acetylglucosamine-1-phosphate transferase, alpha Gnptab 432486 2.65 and beta Mapk1 26413 Mitogen-activated protein kinase 1 2.64 Igf2r 16004 Insulin-like growth factor 2 receptor 2.63 Cysteine-rich secretory protein LCCL domain Crispld2 78892 2.63 containing 2 Ttc3 22129 Tetratricopeptide repeat domain 3 2.61 Cgref1 68567 Cell growth regulator with EF hand domain 1 2.61 Serf1 20365 Small EDRK-rich factor 1 2.60 Ly6c1 17067 Lymphocyte antigen 6 complex, locus C1 2.60 Mkx 210719 Mohawk 2.57 Mmp2** 17390 Matrix metallopeptidase 2 2.57 Reps2 194590 RALBP1 associated Eps domain containing protein 2 2.54 Cyr61 16007 Cysteine rich protein 61 2.54 Xiap 11798 X-linked inhibitor of apoptosis 2.53 Rbbp9 26450 Retinoblastoma binding protein 9 2.52 Mef2a** 17258 Myocyte enhancer factor 2A 2.51 Arl13b 68146 ADP-ribosylation factor-like 13B 2.51 Golga3 269682 Golgi autoantigen, golgin subfamily a, 3 2.49 Csnk2a1 12995 Casein kinase 2, alpha 1 polypeptide 2.48 Dym 69190 Dymeclin 2.46 Vav3 57257 Vav 3 oncogene 2.46 Pftk1 18647 PFTAIRE protein kinase 1 2.41 Lgals2 107753 Lectin, galactose-binding, soluble 2 2.38 Egln1 112405 EGL nine homolog 1 2.38 Bicc1 83675 Bicaudal C homolog 1 2.37 Aldh1a1 11668 Aldehyde dehydrogenase family 1, subfamily A1 2.37 Cand1 71902 Cullin associated and neddylation disassociated 1 2.36 Ndrg4 234593 N-myc downstream regulated gene 4 2.36 Sirtuin 1 (silent mating type information regulation 2, Sirt1 93759 2.34 homolog) 1 Nav1 215690 Neuron navigator 1 2.32 Ghitm 66092 Growth hormone inducible transmembrane protein 2.31 Lgals7 16858 Lectin, galactose binding, soluble 7 2.31 Dph3 105638 DPH3 homolog (KTI11) 2.30 Ube2b 22210 Ubiquitin-conjugating enzyme E2B, RAD6 homolog 2.28

- 147- Appendix

Grem1 23892 Gremlin 1 2.27 Wwp1 107568 WW domain containing E3 ubiquitin protein ligase 1 2.26 Foxc1 17300 Forkhead box C1 2.25 Strn 268980 Striatin, calmodulin binding protein 2.24 Ptprg 19270 Protein tyrosine phosphatase, receptor type, G 2.22 Jph1 57339 Junctophilin 1 2.22 Cyr61 16007 Cysteine rich protein 61 2.21 Fgf18** 14172 Fibroblast growth factor 18 2.21 Grem2 23893 Gremlin 2 homolog, cysteine knot superfamily 2.20 Wnt5a 22418 Wingless-related MMTV integration site 5A 2.19 Edf1 59022 Endothelial differentiation-related factor 1 2.19 Plec1 18810 Plectin 1 2.18 Dmp1 13406 Dentin matrix protein 1 2.18 Serf2 378702 Small EDRK-rich factor 2 2.17 Nrp1** 18186 Neuropilin 1 2.16 Mtap1b** 17755 Microtubule-associated protein 1B 2.14 Mbnl1 56758 Muscleblind-like 1 2.14 Braf 109880 Braf transforming gene 2.13 Stoml2 66592 Stomatin (Epb7.2)-like 2 2.11 Adcyap1r1 11517 Adenylate cyclase activating polypeptide 1 receptor 1 2.11 Cfl2 12632 Cofilin 2, muscle 2.09 Csda 56449 Cold shock domain protein A 2.09 Hsd17b10 15108 Hydroxysteroid (17-beta) dehydrogenase 10 2.08 Herpud2 80517 HERPUD family member 2 2.07 Core 1 synthase, glycoprotein-N- C1galt1 94192 2.06 acetylgalactosamine, 1 Gpm6b 14758 Glycoprotein m6b 2.06 Fbn1 14118 Fibrillin 1 2.06 Ndrg3 29812 N-myc downstream regulated gene 3 2.06 Pak3 18481 P21 (CDKN1A)-activated kinase 3 2.04 Zfp521 225207 Zinc finger protein 521 2.03 Acvr2a** 11480 Activin receptor IIA 2.03 Chad** 12643 Chondroadherin 2.02 Lect1** 16840 Leukocyte cell derived chemotaxin 1 2.01 Mef2c** 17260 Myocyte enhancer factor 2C 2.00 *p value smaller than 0.05 **Genes reportedly associated with bone metabolism

Apoptosis Clu 12759 Clusterin 3.82* Pleckstrin homology-like domain, family A, member Phlda1 21664 3.74 1 Dnm1l 74006 Dynamin 1-like 3.38

- 148- Appendix

Anxa8 11752 Annexin A8 3.28 Sbsn 282619 Suprabasin 3.27 Cdr2 12585 Cerebellar degeneration-related 2 2.99 Ank2 109676 Ankyrin 2, brain 2.94 Ypel2 77864 Yippee-like 2 2.91 Plxdc2 67448 Plexin domain containing 2 2.87 Lrp1 16971 Low density lipoprotein receptor-related protein 1 2.82 Ccdc46 76380 Coiled-coil domain containing 46 2.69 Dleu2 328425 Deleted in lymphocytic leukemia, 2 2.65 Myeloid/lymphoid or mixed-lineage leukemia, Mllt11 56772 2.63 translocated, 11 Klhl7 52323 Kelch-like 7 2.57 Txn2 56551 Thioredoxin 2 2.57 Tia1 21841 Cytotoxic granule-associated RNA binding protein 1 2.52 Use1 67023 Unconventional SNARE in the ER 1 homolog 2.44 Il7** 16196 Interleukin 7 2.39 Ntn1 18208 Netrin 1 2.38 Ptpn13 19249 Protein tyrosine phosphatase, non-receptor type 13 2.35 Glrx2 69367 Glutaredoxin 2 (thioltransferase) 2.34 Neutral sphingomyelinase (N-SMase) activation Nsmaf 18201 2.32 associated factor Magef1 76222 Melanoma antigen family F, 1 2.28 Armcx3 71703 Armadillo repeat containing, X-linked 3 2.27 Spg21 27965 Spastic paraplegia 21 homolog 2.22 Phosphatidylinositol 3-kinase, catalytic, alpha Pik3ca 18706 2.21 polypeptide Bcl2l1 12048 Bcl2-like 1 2.17 Gpatch4 66614 G patch domain containing 4 2.17 Rusc1 72296 RUN and SH3 domain containing 1 2.16 Rnf13 24017 Ring finger protein 13 2.15 Rnft1 76892 Ring finger protein, transmembrane 1 2.13 Ankrd28 105522 Ankyrin repeat domain 28 2.12 Ctage5 217615 CTAGE family, member 5 2.11 Letmd1 68614 LETM1 domain containing 1 2.09 Membrane protein, palmitoylated 5, subfamily Mpp5 56217 2.08 member 5) Uveal autoantigen with coiled-coil domains and Uaca 72565 2.08 ankyrin repeats Rora 19883 RAR-related orphan receptor alpha 2.08 Becn1 56208 Beclin 1, autophagy related 2.06 Casp3 12367 Caspase 3 2.06 Dynll1 56455 Dynein light chain LC8-type 1 2.05 Dnajb9 27362 DnaJ (Hsp40) homolog, subfamily B, member 9 2.05 Mtus1 102103 Mitochondrial tumor suppressor 1 2.04

- 149- Appendix

Col4a2 12827 Collagen, type IV, alpha 2 2.02 Tmem85 68032 Transmembrane protein 85 2.01 Sh3kbp1 58194 SH3-domain kinase binding protein 1 2.00 *p value smaller than 0.05 **Genes reportedly associated with bone metabolism

Cell cycle Mbp 17196 Myelin basic protein 15.07* Adk 11534 Adenosine kinase 5.06 Pkia 18767 Protein kinase inhibitor, alpha 4.16 Smc2 14211 Structural maintenance of chromosomes 2 4.01 Cables1 63955 Cdk5 and Abl enzyme substrate 1 3.35 Ccdc16 66983 Coiled-coil domain containing 16 3.28 NIMA (never in mitosis gene a)-related expressed Nek1 18004 3.15 kinase 1 Cdc23 52563 CDC23 (cell division cycle 23, yeast, homolog) 2.94 Id4 15904 Inhibitor of DNA binding 4 2.92 Gspt1 14852 G1 to S phase transition 1 2.89 Nipbl 71175 Nipped-B homolog 2.75 Pfdn1 67199 Prefoldin 1 2.72 Eea1 216238 Early endosome antigen 1 2.70 Nuclear fragile X mental retardation protein Nufip2 68564 2.59 interacting protein 2 Ncapg 54392 On-SMC condensin I complex, subunit G 2.55 Mtmr6 219135 Myotubularin related protein 6 2.47 Anapc2 99152 Anaphase promoting complex subunit 2 2.47 Rrm2 20135 Ribonucleotide reductase M2 2.38 Calml4 75600 Calmodulin-like 4 2.36 Pole 18973 Polymerase (DNA directed), epsilon 2.33 Calcium/calmodulin-dependent protein kinase II Camk2a 12322 2.29 alpha Clec11a 20256 C-type lectin domain family 11, member a 2.28 Cep76 225659 Centrosomal protein 76 2.26 REV3-like, catalytic subunit of DNA polymerase Rev3l 19714 2.25 zeta RAD54 like Nbn 27354 Nibrin 2.24 Gas5 14455 Growth arrest specific 5 2.22 Dtymk 21915 Deoxythymidylate kinase 2.21 Smc1a 24061 Structural maintenance of chromosomes 1A 2.20 Ccnd2 12444 Cyclin D2 2.20 Bola2 66162 BolA-like 2 (E. coli) 2.19 Cep68 216543 Centrosomal protein 68 2.17 Lrrcc1 71710 Leucine rich repeat and coiled-coil domain contain. 1 2.17

- 150- Appendix

Ccni 12453 Cyclin I 2.14 Cul5 75717 Cullin 5 2.11 Polk 27015 Polymerase (DNA directed), kappa 2.11 Sept7 235072 Septin 7 2.10 Egfr 13649 Epidermal growth factor receptor 2.10 Nuclear autoantigenic sperm protein (histone- Nasp 50927 2.09 binding) NIMA (never in mitosis gene a)-related expressed Nek1 18004 2.07 kinase 1 Pdgfd 71785 Platelet-derived growth factor, D polypeptide 2.03 COP9 (constitutive photomorphogenic) homolog, Cops5 26754 2.03 subunit 5 Foxn3 71375 Forkhead box N3 2.02 Pmp22 18858 Peripheral myelin protein 2.02 Sept5 18951 Septin 5 2.02 Mphosph8 75339 M-phase phosphoprotein 8 2.02 Polymerase (DNA-directed), delta 3, accessory Pold3 67967 2.02 subunit Seh1l 72124 SEH1-like 2.01 *p value smaller than 0.05

Signal transduction Cldnd1 224250 Claudin domain containing 1 6.72* Stat5a 20850 Signal transducer and activator of transcription 5A 5.95 Ptgfr 19220 Prostaglandin F receptor 4.49 Phosphodiesterase 4D interacting protein Pde4dip 83679 4.15 (myomegalin) Pde7b 29863 Phosphodiesterase 7B 4.06 Mfap5 50530 Microfibrillar associated protein 5 4.01 Twf2 23999 Twinfilin, actin-binding protein, homolog 2 3.92 Wasf2 242687 WAS protein family, member 2 3.16 ATPase, Ca++ transporting, cardiac muscle, slow Atp2a2 11938 3.15 twitch 2 Asb13 142688 Ankyrin repeat and SOCS box-containing protein 13 3.13 Tmem55a 72519 Transmembrane protein 55A 3.04 Stk39 53416 Serine/threonine kinase 39, STE20/SPS1 homolog 3.02 Plcb1 18795 Phospholipase C, beta 1 2.82 Riok2 67045 RIO kinase 2 2.82 Rgs5 19737 Regulator of G-protein signaling 5 2.76 Socs4 67296 Suppressor of cytokine signaling 4 2.66 Ulk2 29869 Unc-51 like kinase 2 2.65 Socs6 54607 Suppressor of cytokine signaling 6 2.61 Protein kinase, cAMP dependent regulatory, type II Prkar2b 19088 2.59 beta

- 151- Appendix Endothelial differentiation, lysophosphatidic acid G- Edg2 14745 2.59 protein rec., 2 BC060632 244654 cDNA sequence BC060632 2.55 Mapkap1 227743 Mitogen-activated protein kinase associated protein 1 2.50 Rangap1 19387 RAN GTPase activating protein 1 2.46 Socs2 216233 Suppressor of cytokine signaling 2 2.43 Arhgap12 75415 Rho GTPase activating protein 12 2.41 Trap1 68015 TNF receptor-associated protein 1 2.41 Map2k1 26395 Mitogen-activated protein kinase kinase 1 2.40 Tex2 21763 Testis expressed gene 2 2.39 Epha3 13837 Eph receptor A3 2.37 Cdc42 12540 Cell division cycle 42 homolog 2.35 Rab32 67844 RAB32, member RAS oncogene family 2.33 Rps6kc1 320119 Ribosomal protein S6 kinase polypeptide 1 2.29 Rabif 98710 RAB interacting factor 2.27 Pde5a 242202 Phosphodiesterase 5A, cGMP-specific 2.27 Map3k2 26405 Mitogen-activated protein kinase kinase kinase 2 2.25 Mpp1 17524 Membrane protein, palmitoylated 2.24 Ptpdc1 218232 Protein tyrosine phosphatase domain containing 1 2.23 Mn1** 433938 Meningioma 1 2.23 Rhoj 80837 Ras homolog gene family, member J 2.23 Rcan1 54720 Regulator of calcineurin 1 2.22 Nenf 66208 Neuron derived neurotrophic factor 2.17 Rapgef6 192786 Rap guanine nucleotide exchange factor (GEF) 6 2.16 Itsn1 16443 Intersectin 1 (SH3 domain protein 1A) 2.16 Arf6 11845 ADP-ribosylation factor 6 2.13 Pde1a 18573 Phosphodiesterase 1A, calmodulin-dependent 2.11 Endothelial differentiation, sphingolipid G-protein- Edg3 13610 2.11 coupled rec., 3 Lrba 80877 LPS-responsive beige-like anchor 2.10 Asb4 65255 Ankyrin repeat and SOCS box-containing protein 4 2.08 Asb8 78541 Ankyrin repeat and SOCS box-containing protein 8 2.07 Grb10 14783 Growth factor receptor bound protein 10 2.06 Cyhr1 54151 Cysteine and histidine rich 1 2.06 V-ral simian leukemia viral oncogene homolog A Rala 56044 2.06 (ras related) Cdc42bpa 226751 Cdc42 binding protein kinase alpha 2.04 Akap9 100986 A kinase (PRKA) anchor protein (yotiao) 9 2.04 Arhgap18 73910 Rho GTPase activating protein 18 2.03 Gapvd1 66691 GTPase activating protein and VPS9 domains 1 2.00 *p value smaller than 0.05 **Genes reportedly associated with bone metabolism

- 152- Appendix

Transcription Apobec2 11811 Apolipoprotein B editing complex 2 6.18* Polr3b 70428 Polymerase (RNA) III (DNA directed) polypeptide B 4.80 Small nuclear RNA activating complex, polypeptide Snapc2 102209 4.76 2 Smyd1 12180 SET and MYND domain containing 1 4.55 Tnni2 21953 Troponin I, skeletal, fast 2 4.24 Zfp295 114565 Zinc finger protein 295 4.01 Endod1 71946 Endonuclease domain containing 1 3.99 Lmo4 16911 LIM domain only 4 3.86 Lrrn1 16979 Leucine rich repeat protein 1, neuronal 3.71 Eya4 14051 Eyes absent 4 homolog 3.68 Papd4 100715 PAP associated domain containing 4 3.66 Zfhx4 80892 Zinc finger homeodomain 4 3.54 Pa2g4 18813 Proliferation-associated 2G4 3.47 Zxda 668171 Zinc finger, X-linked, duplicated A 3.43 Egr3 13655 Early growth response 3 3.42 Hnrnpc 15381 Heterogeneous nuclear ribonucleoprotein C 3.41 Trps1 83925 Trichorhinophalangeal syndrome I 3.38 Lsm2 27756 LSM2 homolog, U6 small nuclear RNA associated 3.29 AK129302 245522 cDNA sequence AK129302 3.28 Suv39h2 64707 Suppressor of variegation 3-9 homolog 2 3.21 Zfp637 232337 Zinc finger protein 637 3.16 Hist2h3c2 97114 Histone cluster 2, H3c2 3.14 Smyd2 226830 SET and MYND domain containing 2 3.12 Zfp260 26466 Zinc finger protein 260 3.03 Lin54 231506 Lin-54 homolog 2.96 Prrx1 18933 Paired related homeobox 1 2.93 Ruvbl2 20174 RuvB-like protein 2 2.93 Hif1a 15251 Hypoxia inducible factor 1, alpha subunit 2.87 Rad52 19365 RAD52 homolog 2.85 Bivm 246229 Basic, immunoglobulin-like variable motif containing 2.83 Nfib 18028 Nuclear factor I/B 2.74 Tead1 21676 TEA domain family member 1 2.73 Ets2 23872 E26 avian leukemia oncogene 2, 3' domain 2.72 Sdpr 20324 Serum deprivation response 2.69 Atf5 107503 Activating transcription factor 5 2.67 E2f6 50496 E2F transcription factor 6 2.63 Zfp60 22718 Zinc finger protein 60 2.60 Sfrs1 110809 Splicing factor, arginine/serine-rich 1 (ASF/SF2) 2.57 Slu7 193116 SLU7 splicing factor homolog 2.57 N6amt2 68043 N-6 adenine-specific DNA methyltransferase 2 2.52

- 153- Appendix

Hnrpab 15384 Heterogeneous nuclear ribonucleoprotein A/B 2.51 Tcfl5 277353 Transcription factor-like 5 (basic helix-loop-helix) 2.48 Prrx2 20204 Paired related homeobox 2 2.48 Pole3 59001 Polymerase (DNA directed), epsilon 3 (p17 subunit) 2.47 Nfic 18029 Nuclear factor I/C 2.46 Hoxb6 15414 Homeo box B6 2.44 Regulator of chromosome condensation and BTB Rcbtb2 105670 2.43 domain 2 Satb1 20230 Special AT-rich sequence binding protein 1 2.41 Tcf4 21413 Transcription factor 4 2.41 Cnot2 72068 CCR4-NOT transcription complex, subunit 2 2.37 Zcchc7 319885 Zinc finger, CCHC domain containing 7 2.37 Trove2 20822 TROVE domain family, member 2 2.35 Nr2c1 22025 Nuclear receptor subfamily 2, group C, member 1 2.35 Zfp160 224585 Zinc finger protein 160 2.35 Ncoa6 56406 Nuclear receptor coactivator 6 2.33 Nuclear factor of activated T-cells, calcineurin- Nfatc4 73181 2.33 dependent 4 Satb2 212712 Special AT-rich sequence binding protein 2 2.32 Sfrs18 66625 Splicing factor, arginine/serine-rich 18 2.31 Zfhx3 11906 Zinc finger homeobox 3 2.28 Pcbd2 72562 Pterin 4 alpha carbinolamine dehydratase 2 2.28 Zbtb41 226470 Zinc finger and BTB domain containing 41 homolog 2.28 Bbx 70508 Bobby sox homolog 2.26 Tshz1 110796 Teashirt zinc finger family member 1 2.24 Sfrs3 20383 Splicing factor, arginine/serine-rich 3 (SRp20) 2.22 Zbtb20 56490 Zinc finger and BTB domain containing 20 2.21 Znhit3 448850 Zinc finger, HIT type 3 2.21 Sertad3 170742 SERTA domain containing 3 2.20 Zfp467 68910 Zinc finger protein 467 2.19 Hist1h3d 319149 Histone cluster 1, H3d 2.18 Tardbp 230908 TAR DNA binding protein 2.18 Pax9 18511 Paired box gene 9 2.17 Rnase4 58809 Ribonuclease, RNase A family 4 2.17 Usp21 30941 Ubiquitin specific peptidase 21 2.16 Nfia 18027 Nuclear factor I/A 2.16 Zfp760 240034 Zinc finger protein 760 2.15 Npat 244879 Nuclear protein in the AT region 2.15 Myeloid/lymphoid or mixed-lineage leukemia, Mllt3 70122 2.15 translocated to, 3 Chd4 107932 Chromodomain helicase DNA binding protein 4 2.14 Hsf2 15500 Heat shock factor 2 2.13 Nf2 18016 Neurofibromatosis 2 2.12

- 154- Appendix

Nfyb 18045 Nuclear transcription factor-Y beta 2.11 Zhx1 22770 Rinc fingers and homeoboxes 1 2.11 Rbm6 19654 RNA binding motif protein 6 2.10 Sfrs2 20382 Splicing factor, arginine/serine-rich 2 (SC-35) 2.10 Six5 20475 Sine oculis-related homeobox 5 homolog 2.10 Zfp423 94187 Zinc finger protein 423 2.09 Ets1 23871 E26 avian leukemia oncogene 1, 5' domain 2.09 Gpatch8 237943 G patch domain containing 8 2.08 Zkscan3 72739 Zinc finger with KRAB and SCAN domains 3 2.07 Mettl9 59052 Methyltransferase like 9 2.06 Jmjd1c 108829 Jumonji domain containing 1C 2.06 Klf12 16597 Kruppel-like factor 12 2.06 Ints6 18130 Integrator complex subunit 6 2.05 Hoxb8 15416 Homeo box B8 2.05 Tada3l 101206 Transcriptional adaptor 3 (NGG1 homolog, yeast) 2.05 X-ray complementing defective repair, Chinese Xrcc6 14375 2.05 hamster cells 6 Hipk1 15257 Homeodomain interacting protein kinase 1 2.04 TAF4A RNA polymerase II, TATA box binding Taf4a 228980 2.04 protein Phtf2 68770 Putative homeodomain transcription factor 2 2.03 Gtf2i 14886 General transcription factor II I 2.03 Mxi1 17859 Max interacting protein 1 2.03 Inhibitor of Bruton agammaglobulinemia tyrosine Ibtk 108837 2.03 kinase Dr1 13486 Down-regulator of transcription 1 2.03 Zfp251 71591 Zinc finger protein 251 2.03 Ndn 17984 Necdin 2.03 Snrpd3 67332 Small nuclear ribonucleoprotein D3 2.02 Carm1 59035 Coactivator-associated arginine methyltransferase 1 2.02 Sin3a 20466 Transcriptional regulator, SIN3A 2.02 Strap 20901 Serine/threonine kinase receptor associated protein 2.02 Ring1 19763 Ring finger protein 1 2.02 Arrdc4 66412 Arrestin domain containing 4 2.01 Creb3l2 208647 cAMP responsive element binding protein 3-like 2 2.01 Sirtuin 7 (silent mating type information regulation 2, Sirt7 209011 2.01 homolog) 7 Trip4 56404 Thyroid hormone receptor interactor 4 2.01 *p value smaller than 0.05

Immune response Prg4 96875 Proteoglycan 4 (megakaryocyte stimulating factor) 11.29* Asph 65973 Aspartate-beta-hydroxylase 4.04

- 155- Appendix

Mlf2 30853 Myeloid leukemia factor 2 3.87 Prkca 18750 Protein kinase C, alpha 3.21 Scara3 219151 Scavenger receptor class A, member 3 3.19 Il33 77125 Interleukin 33 3.10 Nope 56741 Neighbor of Punc E11 2.86 Cd55 13136 CD55 antigen 2.68 Tlr4 21898 Toll-like receptor 4 2.59 Nlrx1 270151 NLR family member X1 2.30 Cytochrome P450, family 4, subfamily f, polypeptide Cyp4f16 70101 2.19 16 C2 12263 Complement component 2 (within H-2S) 2.17 Ythdf2 213541 YTH domain family 2 2.01 *p value smaller than 0.05

Mucle contraction Smpx 66106 Small muscle protein, X-linked 8.22* ATPase, Ca++ transporting, cardiac muscle, fast Atp2a1 11937 5.24 twitch 1 Casq1 12372 Calsequestrin 1 5.23 Trdn 76757 Triadin 4.83 Myh7 140781 Myosin, heavy polypeptide 7, cardiac muscle, beta 4.30 Cmya5 76469 Cardiomyopathy associated 5 4.15 Myom2 17930 Myomesin 2 2.34 Myh1 17879 Myosin, heavy polypeptide 1, skeletal muscle, adult 2.27 Hsbp1 68196 Heat shock factor binding protein 1 2.14 Myh6 17888 Myosin, heavy polypeptide 6, cardiac muscle, alpha 2.08 *p value smaller than 0.05

Cytoskeleton remodeling Myoc 17926 Myocilin 7.42* Actn3 11474 Actinin alpha 3 6.27 Lmod2 93677 Leiomodin 2 (cardiac) 4.97 Tpm2 22004 Tropomyosin 2, beta 4.89 Synpo2 118449 Synaptopodin 2 4.45 Actn1 109711 Actinin, alpha 1 4.22 Ank1 11733 Ankyrin 1, erythroid 3.46 Nexn 68810 Nexilin 2.98 Cap2 67252 CAP, adenylate cyclase-associated protein, 2 2.78 Dbnl 13169 Drebrin-like 2.59 Nup98 269966 Nucleoporin 98 2.50 Kif5b 16573 Kinesin family member 5B 2.49 Gm114 228730 Gene model 114, (NCBI) 2.39

- 156- Appendix

Cap1 12331 CAP, adenylate cyclase-associated protein 1 2.33 Ttl 69737 Tubulin tyrosine ligase 2.31 Krt10 16661 Keratin 10 2.28 Kif16b 16558 Kinesin family member 16B 2.28 LysM, putative peptidoglycan-binding, domain Lysmd2 70082 containing 2 2.26 Oxr1 170719 Oxidation resistance 1 2.22 AA536749 26936 Expressed sequence AA536749 2.16 Rdx 19684 Radixin 2.13 Wdr1 22388 WD repeat domain 1 2.13 Wdr46 57315 WD repeat domain 46 2.11 Txndc14 66958 Thioredoxin domain containing 14 2.10 Dmn 233335 Desmuslin 2.10 Sync 68828 Syncoilin 2.08 Tubb6 67951 Tubulin, beta 6 2.08 Actr1b 226977 ARP1 actin-related protein 1 homolog B 2.08 Epb4.1l1 13821 Erythrocyte protein band 4.1-like 1 2.07 Synpo2 118449 Synaptopodin 2 2.07 Dync1h1 13424 Dynein cytoplasmic 1 heavy chain 1 2.06 Sarcoglycan, delta (dystrophin-associated Sgcd 24052 2.05 glycoprotein) Rnf19a 30945 Ring finger protein 19A 2.03 Pfn2 18645 Profilin 2 2.00 *p value smaller than 0.05

Proteolysis Pcolce2 76477 Procollagen C-endopeptidase enhancer 2 3.24* Ariadne ubiquitin-conjugating enzyme binding Arih1 23806 3.10 protein homolog 1 Btbd3 228662 BTB (POZ) domain containing 3 2.89 Prss23 76453 Protease, serine, 23 2.89 Siah1b 20438 Seven in absentia 1B 2.86 Capn1 12333 Calpain 1 2.74 Rnf11 29864 Ring finger protein 11 2.65 Qpctl 67369 Glutaminyl-peptide cyclotransferase-like 2.61 Ddi1 71829 DDI1, DNA-damage inducible 1, homolog 1 2.59 Trip12 14897 Thyroid hormone receptor interactor 12 2.56 Cyld 74256 Cylindromatosis (turban tumor syndrome) 2.54 Dpp7 83768 Dipeptidylpeptidase 7 2.48 Pcolce 18542 Procollagen C-endopeptidase enhancer protein 2.46 Dzip3 224170 DAZ interacting protein 3, zinc finger 2.41 Proteasome (prosome, macropain) 26S subunit, non- Psmd11 69077 2.41 ATPase, 11

- 157- Appendix

Dda1 66498 DET1 and DDB1 associated 1 2.40 Trim37 68729 Tripartite motif-containing 37 2.35 Spop 20747 Speckle-type POZ protein 2.30 Ntan1 18203 N-terminal Asn amidase 2.30 Fbxo33 70611 F-box protein 33 2.27 Usp47 74996 Ubiquitin specific peptidase 47 2.23 Prepl 213760 Prolyl endopeptidase-like 2.22 Pi16 74116 Peptidase inhibitor 16 2.22 Proteasome (prosome, macropain) 26S subunit, Psmc6 67089 2.20 ATPase, 6 Bace1 23821 Beta-site APP cleaving enzyme 1 2.20 Ube3a 22215 Ubiquitin protein ligase E3A 2.13 Ubiquitin-conjugating enzyme E2G 1 (UBC7 Ube2g1 67128 2.12 homolog) Erap1 80898 Endoplasmic reticulum aminopeptidase 1 2.11 Capn2 12334 Calpain 2 2.10 Usp15 14479 Ubiquitin specific peptidase 15 2.09 Usp10 22224 Ubiquitin specific peptidase 10 2.08 Stch 110920 Stress 70 protein chaperone, microsome-associated 2.08 Rnf11 29864 Ring finger protein 11 2.08 A disintegrin and metallopeptidase domain 19 Adam19 11492 2.07 (meltrin beta) Npepps 19155 Aminopeptidase puromycin sensitive 2.05 Kdelc1 72050 KDEL (Lys-Asp-Glu-Leu) containing 1 2.04 Lonp2 66887 Lon peptidase 2, peroxisomal 2.04 Ube3a 22215 Ubiquitin protein ligase E3A 2.04 Fbxl16 214931 F-box and leucine-rich repeat protein 16 2.02 Tulp4 68842 Tubby like protein 4 2.00 *p value smaller than 0.05

Translation Rps20 67427 Ribosomal protein S20 4.61* D10Ertd322e 67270 DNA segment, Chr 10, ERATO Doi 322, expressed 3.71 Pum1 80912 Pumilio 1 3.31 Dus3l 224907 Dihydrouridine synthase 3-like 3.21 Lyrm2 108755 LYR motif containing 2 3.20 Nubp2 26426 Nucleotide binding protein 2 3.13 Sephs2 20768 Selenophosphate synthetase 2 3.08 Rps4y2 66184 Ribosomal protein S4, Y-linked 2 2.93 Ubiquitin-conjugating enzyme E2D 1, UBC4/5 Ube2d1 216080 2.82 homolog DCN1, defective in cullin neddylation 1, domain Dcun1d2 102323 2.62 containing 2

- 158- Appendix

Arl6ip1 54208 ADP-ribosylation factor-like 6 interacting protein 1 2.58 Wars 22375 Tryptophanyl-tRNA synthetase 2.56 Asparaginyl-tRNA synthetase 2 Nars2 244141 2.54 (mitochondrial)(putative) Mrpl15 27395 Mitochondrial ribosomal protein L15 2.53 Hsph1 15505 Heat shock 105kDa/110kDa protein 1 2.53 Eif3h 68135 Eukaryotic translation initiation factor 3, subunit H 2.51 Pus3 67049 Pseudouridine synthase 3 2.46 Ppil3 70225 Peptidylprolyl isomerase (cyclophilin)-like 3 2.45 Pdzd4 245469 PDZ domain containing 4 2.33 Niban 63913 Niban protein 2.32 Stk32b 64293 Serine/threonine kinase 32B 2.30 Mrp63 67840 Mitochondrial ribosomal protein 63 2.30 Selm 114679 Selenoprotein M 2.29 Hspa4 15525 Heat shock protein 4 2.25 Rps27l 67941 Ribosomal protein S27-like 2.22 Ubl4 27643 Ubiquitin-like 4 2.19 Cytoplasmic polyadenylation element binding protein Cpeb3 208922 2.18 3 Rps17 20068 Ribosomal protein S17 2.15 Dph4 99349 DPH4 homolog (JJJ3) 2.15 Mrpl40 18100 Mitochondrial ribosomal protein L40 2.09 Mars 216443 Methionine-tRNA synthetase 2.09 Mrpl43 94067 Mitochondrial ribosomal protein L43 2.08 Qars 97541 Glutaminyl-tRNA synthetase 2.08 Dnajc21 78244 DnaJ (Hsp40) homolog, subfamily C, member 21 2.08 Atad3a 108888 ATPase family, AAA domain containing 3A 2.04 Protein prenyltransferase alpha subunit repeat Ptar1 72351 2.04 containing 1 Tufm 233870 Tu translation elongation factor, mitochondrial 2.03 Eif2a 229317 Eukaryotic translation initiation factor 2a 2.01 *p value smaller than 0.05

G-protein signaling Kcnk2 16526 Potassium channel, subfamily K, member 2 3.38* Atp1b1 11931 ATPase, Na+/K+ transporting, beta 1 polypeptide 3.20 Sphk2 56632 Sphingosine kinase 2 2.45 Git2 26431 G protein-coupled receptor kinase-interactor 2 2.26 Gtpbp2 56055 GTP binding protein 2 2.19 *p value smaller than 0.05

Cell adhesion

- 159- Appendix

Tmem107 66910 Transmembrane protein 107 6.01* Vcan 13003 Versican 5.86 Lyve1 114332 Lymphatic vessel endothelial hyaluronan receptor 1 5.42 Hapln1 12950 Hyaluronan and proteoglycan link protein 1 4.39 Col14a1 12818 Collagen, type XIV, alpha 1 4.20 Tmem157 67698 Transmembrane protein 157 4.16 Hfe2 69585 Hemochromatosis type 2 (juvenile) 3.91 Pcdh7 54216 Protocadherin 7 3.59 Boc 117606 Biregional cell adhesion molecule-related 3.55 Dock4 238130 Dedicator of cytokinesis 4 3.39 LIM domain containing preferred translocation Lpp 210126 3.37 partner in lipoma Fermt3 108101 Fermitin family homolog 3 3.34 Fndc1 68655 Fibronectin type III domain containing 1 3.07 Cd34 12490 CD34 antigen 3.03 Thbs2 21826 Thrombospondin 2 2.97 Itgbl1 223272 Integrin, beta-like 1 2.93 Thbs1** 21825 Thrombospondin 1 2.93 Sspn 16651 Sarcospan 2.90 Pcdhb14 93885 Protocadherin beta 14 2.81 Flrt2 399558 Fibronectin leucine rich transmembrane protein 2 2.76 Abi3bp 320712 ABI gene family, member 3 (NESH) binding protein 2.71 Cell adhesion molecule-related/down-regulated by Cdon 57810 2.65 oncogenes Fbln1 14114 Fibulin 1 2.60 Dpt 56429 Dermatopontin 2.53 Col6a3 12835 Collagen, type VI, alpha 3 2.48 Tmem87b 72477 Transmembrane protein 87B 2.46 Ptprk 19272 Protein tyrosine phosphatase, receptor type, K 2.46 Cd151 12476 CD151 antigen 2.41 Fn1** 14268 Fibronectin 1 2.37 Col6a2 12834 Collagen, type VI, alpha 2 2.36 Mpp7 75739 Membrane protein, palmitoylated 7, member 7 2.35 Lamb2 16779 Laminin, beta 2 2.35 Ptprf 19268 Protein tyrosine phosphatase, receptor type, F 2.31 Efs 13644 Embryonal Fyn-associated substrate 2.30 Lims1 110829 LIM and senescent cell antigen-like domains 1 2.26 Tmem119 231633 Transmembrane protein 119 2.23 Col27a1 373864 Collagen, type XXVII, alpha 1 2.18 Jup 16480 Junction plakoglobin 2.17 Cercam 99151 Cerebral endothelial cell adhesion molecule 2.17 Itga6 16403 Integrin alpha 6 2.12 Comp 12845 Cartilage oligomeric matrix protein 2.07

- 160- Appendix

Tpbg 21983 Trophoblast glycoprotein 2.07 Panx3** 208098 Pannexin 3 2.06 Pcdhb9 93880 Protocadherin beta 9 2.05 Alcam 11658 Activated leukocyte cell adhesion molecule 2.05 Clstn1 65945 Calsyntenin 1 2.03 Thbd 21824 Thrombomodulin 2.02 Pcnx 54604 Pecanex homolog 2.00 *p value smaller than 0.05 **Genes reportedly associated with bone metabolism

Transport Ipo4 75751 Importin 4 7.22* Golt1b 66964 Golgi transport 1 homolog B 5.36 Cox8b 12869 Cytochrome c oxidase, subunit VIIIb 4.40 Ryr1 20190 Ryanodine receptor 1, skeletal muscle 4.19 Transient receptor potential cation channel, Trpc6 22068 4.17 subfamily C, 6 Rab6b 270192 RAB6B, member RAS oncogene family 4.09 Mfap3l 71306 Microfibrillar-associated protein 3-like 3.97 Chsy3 78923 Chondroitin sulfate synthase 3 3.73 B3galnt2 97884 UDP-GalNAc:betaGlcNAc polypeptide 2 3.71 Ssr4 20832 Signal sequence receptor, delta 3.24 Kbtbd7 211255 Kelch repeat and BTB (POZ) domain containing 7 3.21 Myo5a 17918 Myosin Va 3.14 Srpk2 20817 Serine/arginine-rich protein specific kinase 2 3.12 Ttc30b 72421 Tetratricopeptide repeat domain 30B 3.09 NADH dehydrogenase (ubiquinone) 1 alpha Ndufa11 69875 3.09 subcomplex 11 Tmem86a 67893 Transmembrane protein 86A 3.02 Ap3m1 55946 Adaptor-related protein complex 3, mu 1 subunit 2.99 Srpr 67398 Signal recognition particle receptor 2.99 Slc38a10 72055 Solute carrier family 38, member 10 2.99 Intraflagellar transport 122 homolog Ift122 81896 2.93 (Chlamydomonas) Potassium large calcium-activated channel, alpha Kcnma1 16531 2.89 member 1 Kif1b 16561 Kinesin family member 1B 2.85 Cyb5b 66427 Cytochrome b5 type B 2.77 Hook3 320191 Hook homolog 3 2.77 Snx6 72183 Sorting nexin 6 2.74 Stx6 58244 Syntaxin 6 2.73 Sec24a 77371 SEC24 related gene family, member A 2.70 Actr1b 226977 ARP1 actin-related protein 1 homolog B 2.68

- 161- Appendix

Ap2b1 71770 Adaptor-related protein complex 2, beta 1 subunit 2.67 Translocase of outer mitochondrial membrane 70 Tomm70a 28185 2.67 homolog A Mgea5 76055 Meningioma expressed antigen 5 (hyaluronidase) 2.63 Slc35a4 67843 Solute carrier family 35, member A4 2.62 NADH dehydrogenase (ubiquinone) 1 alpha Ndufa7 66416 2.60 subcomplex, 7 Atp1a2 98660 ATPase, Na+/K+ transporting, alpha 2 polypeptide 2.59 Scfd1 76983 Sec1 family domain containing 1 2.58 Stxbp5 78808 Syntaxin binding protein 5 (tomosyn) 2.58 Tpcn1 252972 Two pore channel 1 2.53 Atp1b1 11931 ATPase, Na+/K+ transporting, beta 1 polypeptide 2.52 Protein phosphatase 1, regulatory (inhibitor) subunit Ppp1r12a 17931 2.51 12A Ndufs8 225887 NADH dehydrogenase (ubiquinone) Fe-S protein 8 2.50 Myo1c 17913 Myosin IC 2.48 Kpnb1 16211 Karyopherin (importin) beta 1 2.41 Stau2 29819 Staufen (RNA binding protein) homolog 2 2.40 Uqcrh 66576 Ubiquinol-cytochrome c reductase hinge protein 2.38 ATP synthase, H+ transporting mitochondrial F1 Atp5b 11947 2.37 complex, beta Mobkl3 19070 MOB1, Mps One Binder kinase activator-like 3 2.36 Necap1 67602 NECAP endocytosis associated 1 2.36 Timm17b 21855 Translocase of inner mitochondrial membrane 17b 2.35 Dnaja2 56445 DnaJ (Hsp40) homolog, subfamily A, member 2 2.34 Arfgap2 77038 ADP-ribosylation factor GTPase activating protein 2 2.34 Slc25a4 11739 Solute carrier family 25, member 4 2.32 Vps52 224705 Vacuolar protein sorting 52 2.30 Kif1b 16561 Kinesin family member 1B 2.29 Txndc13 52837 Thioredoxin domain containing 13 2.29 Snapin 20615 SNAP-associated protein 2.28 Slc25a24 229731 Solute carrier family 25, member 24 2.27 Hcfc1r1 353502 Host cell factor C1 regulator 1 (XPO1-dependent) 2.27 Translocase of outer mitochondrial membrane 40 Tomm40 53333 2.25 homolog Stxbp6 217517 Syntaxin binding protein 6 (amisyn) 2.25 Clpb 20480 ClpB caseinolytic peptidase B homolog (E. coli) 2.22 Cnnm3 94218 Cyclin M3 2.21 Pleckstrin homology domain containing, family A Plekha8 231999 2.21 member 8 Bet1 12068 Blocked early in transport 1 homolog 2.19 Sytl4 27359 Bynaptotagmin-like 4 2.18 Arfgap3 66251 ADP-ribosylation factor GTPase activating protein 3 2.16 Vps11 71732 Vacuolar protein sorting 11 2.15

- 162- Appendix

Sft2d2 108735 SFT2 domain containing 2 2.15 StAR-related lipid transfer (START) domain Stard4 170459 2.15 containing 4 Ndufv1 17995 NADH dehydrogenase (ubiquinone) flavoprotein 1 2.14 Tnpo3 320938 Transportin 3 2.14 Slc25a17 20524 Solute carrier family 25, member 17 2.14 Dnajc19 67713 DnaJ (Hsp40) homolog, subfamily C, member 19 2.13 Rufy2 70432 RUN and FYVE domain-containing 2 2.13 Golga4 54214 Golgi autoantigen, golgin subfamily a, 4 2.12 Nup62 18226 Nucleoporin 62 2.11 Exoc4 20336 Exocyst complex component 4 2.10 Solute carrier family 25 (mitochondrial carrier, Slc25a12 78830 2.10 Aralar), member 12 Pleckstrin homology domain containing, family B Plekhb2 226971 2.09 member 2 Scyl2 213326 SCY1-like 2 (S. cerevisiae) 2.08 Slc38a9 268706 Solute carrier family 38, member 9 2.08 Stx18 71116 Syntaxin 18 2.07 Cd320 54219 CD320 antigen 2.06 ATP synthase, H+ transport, mitochondrial F1 Atp5o 28080 2.05 complex, O subunit Ergic3 66366 ERGIC and Golgi 3 2.05 Dscr3 13185 Down syndrome critical region gene 3 2.04 MRS2 magnesium homeostasis factor homolog (S. Mrs2 380836 2.03 cerevisiae) NADH dehydrogenase (ubiquinone) 1 alpha Ndufa5 68202 2.03 subcomplex, 5 Cope 59042 Coatomer protein complex, subunit epsilon 2.02 Ccs 12460 Copper chaperone for superoxide dismutase 2.02 Txnl1 53382 Thioredoxin-like 1 2.02 Sec61a1 53421 Sec61 alpha 1 subunit 2.01 Calcium channel, voltage-dependent, alpha2/delta Cacna2d1 12293 2.01 subunit 1 Snx12 55988 Sorting nexin 12 2.01 Vbp1 22327 Von Hippel-Lindau binding protein 1 2.01 Tmed10 68581 Transmembrane emp24-like trafficking protein 10 2.00 *p value smaller than 0.05

Energy metabolism Coch 12810 Coagulation factor C homolog (Limulus polyphemus) 7.74* Pgam2 56012 Phosphoglycerate mutase 2 6.29 Phkb 102093 Phosphorylase kinase beta 4.65 Pvalb 19293 Parvalbumin 4.54 Acpl2 235534 Acid phosphatase-like 2 4.21

- 163- Appendix

Srl 106393 Sarcalumenin 3.96 Adssl1 11565 Adenylosuccinate synthetase like 1 3.86 AW548124 106522 Expressed sequence AW548124 3.83 Akr1c14 105387 Aldo-keto reductase family 1, member C14 3.77 Acaa2 52538 Acetyl-Coenzyme A acyltransferase 2 3.69 Scarb1 20778 Scavenger receptor class B, member 1 3.62 Art1 11870 ADP-ribosyltransferase 1 3.35 AU RNA binding protein/enoyl-coenzyme A Auh 11992 3.29 hydratase Lias 79464 Lipoic acid synthetase 3.22 Ankrd49 56503 Ankyrin repeat domain 49 3.06 Hagh 14651 Hydroxyacyl glutathione hydrolase 3.02 Dtymk 21915 Deoxythymidylate kinase 2.97 UDP-GlcNAc:betaGal beta-1,3-N- B3gnt1 108902 2.93 acetylglucosaminyltransferase 1 Cmbl 69574 Carboxymethylenebutenolidase-like (Pseudomonas) 2.92 Vldlr 22359 Very low density lipoprotein receptor 2.88 Glutamic pyruvate transaminase (alanine Gpt2 108682 2.88 aminotransferase) 2 Dguok 27369 Deoxyguanosine kinase 2.76 Dlat 235339 Dihydrolipoamide S-acetyltransferase 2.74 Loxl4 67573 Lysyl oxidase-like 4 2.71 Aco2 11429 Aconitase 2, mitochondrial 2.68 Gatm 67092 Glycine amidinotransferase (L-arginine:glycine) 2.67 Ampd1 229665 Adenosine monophosphate deaminase 1 (isoform M) 2.65 Tpi1 21991 Triosephosphate isomerase 1 2.60 Acsl3 74205 Acyl-CoA synthetase long-chain family member 3 2.56 Gmds 218138 GDP-mannose 4, 6-dehydratase 2.55 Uck2 80914 Uridine-cytidine kinase 2 2.54 SMEK homolog 2, suppressor of mek1 Smek2 104570 2.54 (Dictyostelium) Shmt2 108037 Serine hydroxymethyltransferase 2 (mitochondrial) 2.53 Gnpda2 67980 Glucosamine-6-phosphate deaminase 2 2.45 Nmral1 67824 NmrA-like family domain containing 1 2.45 Tmem195 319660 Transmembrane protein 195 2.44 Gbe1 74185 Glucan (1,4-alpha-), branching enzyme 1 2.44 Pex19 19298 Peroxisome biogenesis factor 19 2.42 Paics 67054 Phosphoribosylaminoimidazole carboxylase 2.39 Srd5a3 57357 Steroid 5 alpha-reductase 3 2.39 Cpt1a 12894 Carnitine palmitoyltransferase 1a, liver 2.39 Gnpnat1 54342 Glucosamine-phosphate N-acetyltransferase 1 2.37 Sgms2 74442 Sphingomyelin synthase 2 2.36 B3galnt2 97884 UDP-GalNAc:betaGlcNAc beta, polypeptide 2 2.36

- 164- Appendix

Gpx7 67305 Glutathione peroxidase 7 2.34 Ndufv1 17995 NADH dehydrogenase (ubiquinone) flavoprotein 1 2.32 Hsd17b4 15488 Hydroxysteroid (17-beta) dehydrogenase 4 2.27 Hadh 15107 Hydroxyacyl-Coenzyme A dehydrogenase 2.25 Cyb5r3 109754 Cytochrome b5 reductase 3 2.25 Ampd3 11717 AMP deaminase 3 2.23 Cyb5r1 72017 Cytochrome b5 reductase 1 2.23 Ggta1 14594 Glycoprotein galactosyltransferase alpha 1, 3 2.21 Phosphatidylinositol glycan anchor biosynthesis, Pigp 56176 2.21 class P Srd5a3 57357 Steroid 5 alpha-reductase 3 2.21 Ptdss2 27388 Phosphatidylserine synthase 2 2.20 Oxnad1 218885 Oxidoreductase NAD-binding domain containing 1 2.17 Pyridoxal-dependent decarboxylase domain Pdxdc1 94184 2.14 containing 1 Acad9 229211 Acyl-Coenzyme A dehydrogenase family, member 9 2.14 Serinc1 56442 Serine incorporator 1 2.14 Atp10a 11982 ATPase, class V, type 10A 2.14 Hyou1 12282 Hypoxia up-regulated 1 2.13 Pofut2 80294 Protein O-fucosyltransferase 2 2.13 Cpt2 12896 Carnitine palmitoyltransferase 2 2.13 Sc4mol 66234 Sterol-C4-methyl oxidase-like 2.12 Uxs1 67883 UDP-glucuronate decarboxylase 1 2.12 Enpp1 18605 Ectonucleotide pyrophosphatase/phosphodiesterase 1 2.12 Pgam5 72542 Phosphoglycerate mutase family member 5 2.11 Inpp5f 101490 Inositol polyphosphate-5-phosphatase F 2.11 Coiled-coil-helix-coiled-coil-helix domain containing Chchd3 66075 2.11 3 Phyh 16922 Phytanoyl-CoA hydroxylase 2.10 Sod3 20657 Superoxide dismutase 3, extracellular 2.09 Uap1 107652 UDP-N-acetylglucosamine pyrophosphorylase 1 2.09 Ugp2 216558 UDP-glucose pyrophosphorylase 2 2.08 Daglb 231871 Diacylglycerol lipase, beta 2.07 Gss 14854 Glutathione synthetase 2.07 Gyg 27357 Glycogenin 2.07 Gsto1 14873 Glutathione S-transferase omega 1 2.06 Aldoa 11674 Aldolase 1, A isoform 2.06 Phosphoenolpyruvate carboxykinase 2 Pck2 74551 2.06 (mitochondrial) Gstm5 14866 Glutathione S-transferase, mu 5 2.06 Glud1 14661 Glutamate dehydrogenase 1 2.05 Csad 246277 Cysteine sulfinic acid decarboxylase 2.03 Branched chain ketoacid dehydrogenase E1, beta Bckdhb 12040 2.03 polypeptide

- 165- Appendix

Ggps1 14593 Geranylgeranyl diphosphate synthase 1 2.02 Fnip1 216742 Folliculin interacting protein 1 2.00 *p value smaller than 0.05

Neurophysiological process Txlnb 378431 Taxilin beta 6.93* Samd9l 209086 Sterile alpha motif domain containing 9-like 3.12 Ubqln1 56085 Ubiquilin 1 2.96 Maob 109731 Monoamine oxidase B 2.92 Trim2 80890 Tripartite motif-containing 2 2.48 Cltb 74325 Clathrin, light polypeptide (Lcb) 2.46 Wdr21 73828 WD repeat domain 21 2.31 Lrrtm4 243499 Leucine rich repeat transmembrane neuronal 4 2.29 Bpnt1 23827 Bisphosphate 3'-nucleotidase 1 2.03 *p value smaller than 0.05

Unknown process 4-nitrophenylphosphatase and SNAP25-like protein Nipsnap1 18082 5.05* homolog 1 BB125219 105063 Expressed sequence BB125219 3.04 AI843755 100215 Expressed sequence AI843755 3.01 D7Ertd183e 52234 DNA segment, Chr 7, ERATO Doi 183, expressed 2.89 Uhrf1bp1l 75089 UHRF1 (ICBP90) binding protein 1-like 2.89 D7Ertd183e 52234 DNA segment, Chr 7, ERATO Doi 183, expressed 2.78 Auts2 319974 Autism susceptibility candidate 2 2.64 Heatr5b 320473 HEAT repeat containing 5B 2.57 BC034902 228642 cDNA sequence BC034902 2.45 Ttc19 72795 Tetratricopeptide repeat domain 19 2.43 Ssna1 68475 Sjogren's syndrome nuclear autoantigen 1 2.42 Rspry1 67610 Ring finger and SPRY domain containing 1 2.42 Ankrd42 73845 Ankyrin repeat domain 42 2.40 Ehbp1l1 114601 EH domain binding protein 1-like 1 2.39 Ccdc124 234388 Coiled-coil domain containing 124 2.38 D6Ertd474e 52285 DNA segment, Chr 6, ERATO Doi 474, expressed 2.38 Lrrc42 77809 Leucine rich repeat containing 42 2.35 Ccdc3 74186 Coiled-coil domain containing 3 2.31 Bola3 78653 BolA-like 3 2.24 Lrrc45 217366 Leucine rich repeat containing 45 2.22 D10Ertd641e 52717 DNA segment, Chr 10, ERATO Doi 641, expressed 2.20 Prr16 71373 Proline rich 16 2.19 Ankrd50 99696 Ankrin repeat domain 50 2.19 BC046331 230967 cDNA sequence BC046331 2.18

- 166- Appendix

Krcc1 57896 Lysine-rich coiled-coil 1 2.14 AI597468 103266 Expressed sequence AI597468 2.13 D10Ertd641e 52717 DNA segment, Chr 10, ERATO Doi 641, expressed 2.09 Deb1 26901 Differentially expressed in B16F101 2.06 AI503316 105860 Expressed sequence AI503316 2.04 Small nucleolar RNA host gene (non-protein coding) Snhg7 72091 2.03 7 Msl2l1 77853 Male-specific lethal 2-like 1 2.02 *p value smaller than 0.05

- 167- Appendix Table A4. Down-regulated gene expression in trabecular osteocytes induced by multiple loading doses

Cell growth and differentiation Entrez Fold Gene Symbol Gene Gene Description Change ID Apcdd1** 494504 Adenomatosis polyposis coli down-regulated 1 6.06* Spermatogenesis assocciate-glutamate (E)-rich Speer4b 73526 5.96 protein 4b Speer8-ps1 74062 Spermatogenesis glutamate (E)-rich protein 8, ps-1 5.55 Speer7-ps1 75858 Spermatogenesis glutamate (E)-rich protein 7, ps-1 5.51 Gmcl1l 71847 Germ cell-less homolog 1 (Drosophila-like) 5.51 Prlr 19116 Prolactin receptor 4.76 Ctnna2 12386 Catenin (cadherin associated protein), alpha 2 4.64 Speer1-ps1 70896 Spermatogenesis glutamate (E)-rich protein 1, ps-1 4.36 Prelid2 77619 PRELI domain containing 2 4.17 Metrnl 210029 Meteorin, glial cell differentiation regulator-like 3.84 Mtm1 17772 X-linked myotubular myopathy gene 1 3.83 Dtx1 14357 Deltex 1 homolog (Drosophila) 3.71 Aff2 14266 AF4/FMR2 family, member 2 3.67 Aplp2 11804 Amyloid beta (A4) precursor-like protein 2 3.63 Onecut1 15379 One cut domain, family member 1 3.54 Rgmb 68799 RGM domain family, member B 3.50 Nrp 654309 Neural regeneration protein 3.40 Srpk3 56504 Serine/arginine-rich protein specific kinase 3 3.38 Tdrd7 100121 Tudor domain containing 7 3.37 Art2a 11871 ADP-ribosyltransferase 2a 3.32 Hdgfl1 15192 Hepatoma derived growth factor-like 1 3.31 Aldh1a1 11668 Aldehyde dehydrogenase family 1, subfamily A1 3.26 Aldh8a1 237320 Aldehyde dehydrogenase 8 family, member A1 3.19 Solute carrier family 22 (organic cation transporter), Slc22a16 70840 3.15 16 Tmem132d 243274 Transmembrane protein 132D 3.12 Iapp** 15874 Islet amyloid polypeptide 3.10 Rhox2a 75199 Reproductive homeobox 2A 3.03 Bzw2 66912 Basic leucine zipper and W2 domains 2 3.00 Rora 19883 RAR-related orphan receptor alpha 2.99 Bbs7 71492 Bardet-Biedl syndrome 7 2.98 UTP20, small subunit processome component, Utp20 70683 2.94 homolog Hemt1 15202 Hematopoietic cell transcript 1 2.93 Cxxc4 319478 CXXC finger 4 2.93 Meg3** 17263 Maternally expressed 3 2.92

- 168- Appendix

Amtn** 71421 Amelotin 2.87 Core-binding factor, runt domain, alpha subunit 2, Cbfa2t2 12396 2.86 transloc., 2 Nell1** 338352 NEL-like 1 2.81 Magea3 17139 Melanoma antigen, family A, 3 2.72 Pdzrn3 55983 PDZ domain containing RING finger 3 2.70 Npn2 18153 Neoplastic progression 2 2.66 Growth factor receptor bound protein 2-associated Gab2** 14389 2.63 protein 2 Ambra1 228361 Autophagy/beclin 1 regulator 1 2.58 Schip1 30953 Schwannomin interacting protein 1 2.57 Rag1 19373 Recombination activating gene 1 2.56 Oog2 381570 Oogenesin 2 2.50 Usmg2 83677 Upregulated during skeletal muscle growth 2 2.46 Tex19 73679 Testis expressed gene 19 2.40 Dym 69190 Dymeclin 2.37 Spata19 75469 Spermatogenesis associated 19 2.36 Mastl 67121 Microtubule associated serine/threonine kinase-like 2.35 Kirrel3 67703 Kin of IRRE like 3 2.32 Dppa4 73693 Developmental pluripotency associated 4 2.30 Sox10 20665 SRY-box containing gene 10 2.29 Gzmn 245839 Granzyme N 2.27 Mxra8 74761 Matrix-remodelling associated 8 2.26 Edar 13608 Ectodysplasin-A receptor 2.24 Strbp 20744 Spermatid perinuclear RNA binding protein 2.22 Lenep 57275 Lens epithelial protein 2.22 Cd53 12508 CD53 antigen 2.21 Prp2 83380 Proline rich protein 2 2.20 Prl7c1 67505 Prolactin family 7, subfamily c, member 1 2.17 Tmc1 13409 Transmembrane channel-like gene family 1 2.15 Tex15 104271 Testis expressed gene 15 2.15 Gsdmc1 83492 Gasdermin C1 2.15 Gpc2 71951 Glypican 2 (cerebroglycan) 2.15 Utrn 22288 Utrophin 2.13 Svs5 20944 Seminal vesicle secretory protein 5 2.12 Ntrk3 18213 Neurotrophic tyrosine kinase, receptor, type 3 2.09 Tgm3 21818 Transglutaminase 3, E polypeptide 2.08 Dppa5a 434423 Developmental pluripotency associated 5A 2.08 Lefty2 320202 Left-right determination factor 2 2.07 Fezf2 54713 Fez family zinc finger 2 2.05 Hmgb3 15354 High mobility group box 3 2.05 Phyhipl 70911 Phytanoyl-coA hydroxylase interacting protein-like 2.03 Pou3f2 18992 POU domain, class 3, transcription factor 2 2.03

- 169- Appendix

Sprr3 20766 Small proline-rich protein 3 2.01 Hs6st2 50786 Heparan sulfate 6-O-sulfotransferase 2 2.01 *p value smaller than 0.05 **Genes reportedly associated with bone metabolism

Apoptosis Alpk2 225638 Alpha-kinase 2 5.22* Aplp1 11803 Amyloid beta (A4) precursor-like protein 1 4.52 Aim1 11630 Absent in melanoma 1 4.49 Pawr 114774 PRKC, apoptosis, WT1, regulator 3.93 Sp110 109032 Sp110 nuclear body protein 3.89 Lrrc4 192198 Leucine rich repeat containing 4 3.05 Lrrc4c 241568 Leucine rich repeat containing 4C 2.98 Bmf 171543 Bcl2 modifying factor 2.94 Ptk2 14083 PTK2 protein tyrosine kinase 2 2.71 Cell death-inducing DNA fragm. factor, α-subunit Cideb 12684 2.64 effector B Lck 16818 Lymphocyte protein tyrosine kinase 2.57 Psen2 19165 Presenilin 2 2.55 Tumor necrosis factor receptor superfamily, member Tnfrsf9 21942 2.51 9 Sh3rf1 59009 SH3 domain containing ring finger 1 2.49 X-ray complement. defective repair, Chinese hamster Xrcc6 14375 2.40 cells 6 Ccnb1ip1 239083 Cyclin B1 interacting protein 1 2.39 Rnf7 19823 Ring finger protein 7 2.27 Rnf216 108086 Ring finger protein 216 2.26 Cckbr 12426 Cholecystokinin B receptor 2.13 Klrg2 74253 Killer cell lectin-like receptor subfamily G, member 2 2.04 Mageh1 75625 Melanoma antigen, family H, 1 2.03 Pim2 18715 Proviral integration site 2 2.02 C6 12274 Complement component 6 2.01 Fastkd2 75619 FAST kinase domains 2 2.01 *p value smaller than 0.05

Cell cycle Ran 19384 RAN, member RAS oncogene family 4.85* Neural precursor cell expressed, down-regulated gene Nedd8 18002 3.47 8 Ubd 24108 Ubiquitin D 3.03 Anapc10 68999 Anaphase promoting complex subunit 10 2.97 Usp9x 22284 Ubiquitin specific peptidase 9, X chromosome 2.91 Phip 83946 Pleckstrin homology domain interacting protein 2.83

- 170- Appendix

Nek7 59125 NIMA-related expressed kinase 7 2.80 Zzef1 195018 Zinc finger, ZZ-type with EF hand domain 1 2.70 Tbc1d12 209478 TBC1D12: TBC1 domain family, member 12 2.65 Mtus1 102103 Mitochondrial tumor suppressor 1 2.64 Aurkc 20871 Aurora kinase C 2.63 Psca 72373 Prostate stem cell antigen 2.53 Cdc14a 229776 CDC14 cell division cycle 14 homolog A 2.51 Dclre1a 55947 DNA cross-link repair 1A, PSO2 homolog 2.49 Cfdp1 23837 Craniofacial development protein 1 2.46 Incenp 16319 Inner centromere protein 2.46 Brwd3 382236 Bromodomain and WD repeat domain containing 3 2.43 Polh 80905 Polymerase (DNA directed), eta (RAD 30 related) 2.40 Lin52 217708 Lin-52 homolog 2.38 Spin2 278240 Spindlin family, member 2 2.37 Trpd52l3 66745 Tumor protein D52-like 3 2.31 Kifc1 16580 Kinesin family member C1 2.31 Pif1 208084 PIF1 5'-to-3' DNA helicase homolog 2.31 Ube2c 68612 Ubiquitin-conjugating enzyme E2C 2.13 Hcfc1 15161 Host cell factor C1 2.12 Cdca7 66953 Cell division cycle associated 7 2.11 Jmjd5 77035 Jumonji domain containing 5 2.10 Sept14 74211 Septin 14 2.07 Msh3 17686 Mut-S homolog 3 (E. coli) 2.07 Dclre1b 140917 DNA cross-link repair 1B, PSO2 homolog 2.06 Hus1 15574 Hus1 homolog (S. pombe) 2.04 Excision cross-complementing rodent repair Ercc4 50505 2.03 deficiency, 4 Dis3l2 208718 DIS3 mitotic control homolog (S. cerevisiae-like) 2 2.00 *p value smaller than 0.05

Signal transduction V1ra3 113845 Vomeronasal 1 receptor, A3 5.33* V1rd14 81011 Vomeronasal 1 receptor, D14 5.21 Npr1 18160 Natriuretic peptide receptor 1 4.67 Embryonic lethal, abnormal vision, Dros-like 2, Hu Elavl2 15569 4.57 ag B Embryonic lethal, abnormal vision, Dros-like 1, Hu Elavl1 15568 3.86 ag R Traf6 22034 Tnf receptor-associated factor 6 3.78 Vmn2r29 76229 Vomeronasal 2, receptor 29 3.73 Tspan3 56434 Tetraspanin 3 3.66 Centg2 347722 Centaurin, gamma 2 3.55 Vmn2r42 22310 Vomeronasal 2, receptor 42 3.53

- 171- Appendix

Olfr2 18317 Olfactory receptor 2 3.51 Sstr3 20607 Somatostatin receptor 3 3.37 Dlg2 23859 Discs, large homolog 2 3.37 Mapk10 26414 Mitogen-activated protein kinase 10 3.34 Rapgefl1 268480 Rap guanine nucleotide exchange factor GEF-like 1 3.28 Adcy1 432530 Adenylate cyclase 1 3.11 Gnao1 14681 Guanine nucleotide binding protein, alpha O 3.05 Cabin1 104248 Calcineurin binding protein 1 3.05 Grk5 14773 G protein-coupled receptor kinase 5 2.99 Olfml2a 241327 Olfactomedin-like 2A 2.92 Rhbdl3 246104 Rhomboid, veinlet-like 3 2.87 Akap3 11642 A kinase (PRKA) anchor protein 3 2.83 Prkcb1 18751 Protein kinase C, beta 1 2.69 Ppp2r5c 26931 Protein phosphatase 2, subunit B, gamma isoform 2.66 Htr1d 15552 5-hydroxytryptamine (serotonin) receptor 1D 2.57 Latent transforming growth factor beta binding Ltbp1 268977 2.45 protein 1 Odz2 23964 Odd Oz/ten-m homolog 2 2.43 Dgkb 217480 Diacylglycerol kinase, beta 2.42 Snx29 74478 Sorting nexin 29 2.41 Ptprn 19275 Protein tyrosine phosphatase, receptor type, N 2.40 Map2k6 26399 Mitogen-activated protein kinase kinase 6 2.37 Gnal 14680 Guanine nucleotide binding protein, α-stimulating 2.37 Cd247 12503 CD247 antigen 2.36 Klk6 19144 Kallikrein related-peptidase 6 2.36 Plcb1 18795 Phospholipase C, beta 1 2.33 Ankrd44 329154 Ankyrin repeat domain 44 2.25 Smyd5 232187 SET and MYND domain containing 5 2.24 Garnl3 99326 GTPase activating RANGAP domain-like 3 2.22 V1rb4 113854 Vomeronasal 1 receptor, B4 2.20 Lag3 16768 Lymphocyte-activation gene 3 2.20 Rab39b 67790 RAB39B, member RAS oncogene family 2.17 Mitogen-activated protein kinase 8 interacting protein Mapk8ip1 19099 2.16 1 Depdc5 277854 DEP domain containing 5 2.14 Ralgps1 241308 Ral GEF with PH domain and SH3 binding motif 1 2.13 Vmn2r1 56544 Vomeronasal 2, receptor 1 2.09 Ilvbl 216136 IlvB (bacterial acetolactate synthase)-like 2.08 Gngt1 14699 Guanine nucleotide binding protein, polypeptide 1 2.02 Lax1 240754 Lymphocyte transmembrane adaptor 1 2.00 *p value smaller than 0.05

Transcription

- 172- Appendix

Hif3a 53417 Hypoxia inducible factor 3, alpha subunit 5.80* Rpp25 102614 Ribonuclease P 25 subunit 5.04 Zhx1 22770 Zinc fingers and homeoboxes 1 4.80 Snord116 64243 Small nucleolar RNA, C/D box 116 4.71 D3Ertd300e 56790 DNA segment, Chr 3, ERATO Doi 300, expressed 4.20 Lztfl1 93730 Leucine zipper transcription factor-like 1 4.18 Zscan22 232878 Zinc finger and SCAN domain containing 22 4.04 Cux2 13048 Cut-like homeobox 2 3.96 Tbx20 57246 T-box 20 3.85 Kri1 215194 KRI1 homolog 3.82 Esrrg 26381 Estrogen-related receptor gamma 3.79 Ott 18422 Ovary testis transcribed 3.62 Myt1 17932 Myelin transcription factor 1 3.58 Tcfap2b 21419 Transcription factor AP-2 beta 3.52 Wiz 22404 Widely-interspaced zinc finger motifs 3.48 Rsbn1 229675 Rosbin, round spermatid basic protein 1 3.42 Ssbp3 72475 Single-stranded DNA binding protein 3 3.40 Cnot8 69125 CCR4-NOT transcription complex, subunit 8 3.38 Kcnq1ot1 63830 KCNQ1 overlapping transcript 1 3.35 Elk3 13713 ELK3, member of ETS oncogene family 3.10 Hist2h2bb 319189 Histone cluster 2, H2bb 2.95 Zswim3 67538 Zinc finger, SWIM domain containing 3 2.90 Pole2 18974 Polymerase (DNA directed), epsilon 2, p59 subunit 2.85 Gm397 245109 Gene model 397, (NCBI) 2.80 Trim71 636931 Tripartite motif-containing 71 2.79 Trim27 19720 Tripartite motif-containing 27 2.74 Shq1 72171 SHQ1 homolog 2.71 Gmeb1 56809 Glucocorticoid modulatory element bind protein 1 2.71 Runt-related transcription factor 1, translocated, 1 Runx1t1 12395 2.70 (cyclin D-rel.) Foxn2 14236 Forkhead box N2 2.70 Hmga2 15364 High mobility group AT-hook 2 2.68 Ints10 70885 Integrator complex subunit 10 2.68 Zfp133 171588 Zinc finger protein 133 2.65 Zfp609 214812 Zinc finger protein 609 2.63 Processing of precursor 1, ribonuclease P/MRP Pop1 67724 2.60 family Helz 78455 Helicase with zinc finger domain 2.59 Ccrn4l 12457 CCR4 carbon catabolite repression 4-like 2.58 Safb 224903 Scaffold attachment factor B 2.58 Zim3 116811 Zinc finger, imprinted 3 2.55 Myeloid/lymphoid leukemia, trithorax homolog, Mllt10 17354 2.54 transloc., 10

- 173- Appendix

Meis2 17536 Meis homeobox 2 2.53 Pbx1 18514 Pre B-cell leukemia transcription factor 1 2.52 Jmjd1c 108829 Jumonji domain containing 1C 2.52 Lmnb1 16906 Lamin B1 2.51 Tyms 22171 Thymidylate synthase 2.51 Entpd7 93685 Ectonucleoside triphosphate diphosphohydrolase 7 2.51 Nr5a1 26423 Nuclear receptor subfamily 5, group A, member 1 2.46 Klf3 16599 Kruppel-like factor 3 (basic) 2.41 Exosc2 227715 Exosome component 2 2.39 Atxn1 20238 Ataxin 1 2.37 Adi1 104923 Acireductone dioxygenase 1 2.35 Prr8 381626 Proline rich 8 2.35 Grhl1 195733 Grainyhead-like 1 2.30 Ascc2 75452 Activating signal cointegrator 1 complex subunit 2 2.28 Lbxcor1 207667 Ladybird homeobox 1 homolog co-repressor 1 2.28 Survivor of motor neuron protein interacting protein Sip1 66603 2.27 1 RNA guanine-9-methyltransferase domain containing Rg9mtd2 108943 2.27 2 Phb2 12034 Prohibitin 2 2.27 Mta3 116871 Metastasis associated 3 2.25 Myeloid/lymphoid leukemia, trithorax homolog, Mllt11 56772 2.24 transloc. 11 Ubiquitously transc. tetratricopeptide repeat gene, X- Utx 22289 2.23 chrom. Polr2a 20020 Polymerase (RNA) II, DNA directed polypeptide A 2.19 Zfp532 328977 Zinc finger protein 532 2.18 AU RNA binding protein/enoyl-coenzyme A Auh 11992 2.17 hydratase Zfp28 22690 Zinc finger protein 28 2.16 Cdx2 12591 Caudal type homeo box 2 2.16 Hnf1b 21410 HNF1 homeobox B 2.14 Rnasen 14000 Ribonuclease III, nuclear 2.12 Zfp191 59057 Zinc finger protein 191 2.11 Foxp4 74123 Forkhead box P4 2.09 Nr2f2 11819 Nuclear receptor subfamily 2, group F, member 2 2.07 Trim46 360213 Tripartite motif-containing 46 2.07 Ppih 66101 Peptidyl prolyl isomerase H 2.06 Rpap2 231571 RNA polymerase II associated protein 2 2.06 Nfe2l3 18025 Nuclear factor, erythroid derived 2, like 3 2.05 Aff3 16764 AF4/FMR2 family, member 3 2.05 Dpf1 29861 D4, zinc and double PHD fingers family 1 2.05 Hif3a 53417 Hypoxia inducible factor 3, alpha subunit 2.05 Gtf3c1 233863 General transcription factor III C 1 2.03

- 174- Appendix

Chd5 269610 Chromodomain helicase DNA binding protein 5 2.03 *p value smaller than 0.05

Immune response Saa4 20211 Serum amyloid A 4 5.13* Psg19 26439 Pregnancy specific glycoprotein 19 3.76 Nlrp4e 446099 NLR family, pyrin domain containing 4E 3.61 Mageb1 17145 Melanoma antigen, family B, 1 3.16 Tcra 21473 T-cell receptor alpha chain 3.04 C9 12279 Complement component 9 3.01 H2-M10.1 14985 Histocompatibility 2, M region locus 10.1 3.01 Il1f5 54450 Interleukin 1 family, member 5 (delta) 2.87 Sectm1b 58210 Secreted and transmembrane 1B 2.62 Gimap7 231932 GTPase, IMAP family member 7 2.51 Clec2h 94071 C-type lectin domain family 2, member h 2.50 Ly6g6c 68468 Lymphocyte antigen 6 complex, locus G6C 2.45 Nlrp9b 243874 NLR family, pyrin domain containing 9B 2.45 Lrrc40 67144 Leucine rich repeat containing 40 2.36 Xrra1 446101 X-ray radiation resistance associated 1 2.31 Il12rb2 16162 Interleukin 12 receptor, beta 2 2.24 Cd8a 12525 CD8 antigen, alpha chain 2.17 H2-T23 15040 Histocompatibility 2, T region locus 23 2.14 Icos 54167 Inducible T-cell co-stimulator 2.11 Ccdc130 67736 Coiled-coil domain containing 130 2.03 Azgp1 12007 Alpha-2-glycoprotein 1, zinc 2.01 *p value smaller than 0.05

Muscle contraction Myh11 17880 Myosin, heavy polypeptide 11, smooth muscle 3.12* Sntb1 20649 Syntrophin, basic 1 2.45 Arl15 218639 ADP-ribosylation factor-like 15 2.20 Sh3bgr 50795 SH3-binding domain glutamic acid-rich protein 2.11 *p value smaller than 0.05

Cytoskeleton remodeling Tmod1 21916 Tropomodulin 1 7.34* Obsl1 98733 Obscurin-like 1 6.69 Krt6b 16688 Keratin 6B 4.75 Coro2b 235431 Coronin, actin binding protein, 2B 4.42 Mid1 17318 midline 1 3.57 Tubgcp5 233276 Tubulin, gamma complex associated protein 5 3.30 Sfi1 78887 Sfi1 homolog, spindle assembly associated 3.17

- 175- Appendix

Hmmr 15366 Hyaluronan mediated motility receptor (RHAMM) 2.95 Limk1 16885 LIM-domain containing, protein kinase 2.90 Ccdc46 76380 Coiled-coil domain containing 46 2.85 Ccdc51 66658 Coiled-coil domain containing 51 2.83 Sprr2a 20755 Small proline-rich protein 2A 2.73 Lrrc50 68270 Leucine rich repeat containing 50 2.57 Ccdc67 234964 Coiled-coil domain containing 67 2.49 Capn10 23830 Calpain 10 2.33 Wasl 73178 Wiskott-Aldrich syndrome-like 2.30 Fuz 70300 Fuzzy homolog 2.30 Evl 14026 Ena-vasodilator stimulated phosphoprotein 2.29 Krt39 237934 Keratin 39 2.29 Tubb4 22153 Tubulin, beta 4 2.27 Kif6 319991 Kinesin family member 6 2.19 Ccdc83 75338 Coiled-coil domain containing 83 2.19 Diap1 13367 Diaphanous homolog 1 2.18 Fcho2 218503 FCH domain only 2 2.14 Scaper 244891 S phase cyclin A-associated protein in the ER 2.11 Tmsb10 19240 Thymosin, beta 10 2.05 *p value smaller than 0.05

Proteolysis HECT, C2 and WW domain E3 ubiquitin protein Hecw1 94253 5.40* ligase 1 Serpinb9g 93806 Serine peptidase inhibitor, clade B, member 9g 4.83 Klk4 56640 Kallikrein related-peptidase 4 4.52 Tmprss2 50528 Transmembrane protease, serine 2 3.24 Pm20d1 212933 Peptidase M20 domain containing 1 3.06 Rbck1 24105 RanBP-type, C3HC4-type zinc finger containing 1 2.89 Mcpt9 17232 Mast cell protease 9 2.68 Wdr66 269701 WD repeat domain 66 2.67 Stfa3 20863 Stefin A3 2.58 Fbxo36 66153 F-box protein 36 2.46 Serpinb9b 20706 Serine peptidase inhibitor, clade B, member 9b 2.44 Ptpn3 545622 Protein tyrosine phosphatase, non-receptor type 3 2.43 Ermp1 226090 Endoplasmic reticulum metallopeptidase 1 2.42 Ube2j2 140499 Ubiquitin-conjugating enzyme E2, J2 homolog 2.37 Rfwd2 26374 Ring finger and WD repeat domain 2 2.35 Adam28 13522 A disintegrin and metallopeptidase domain 28 2.28 Adam1b 280667 A disintegrin and metallopeptidase domain 1b 2.04 Disintegrin-like/metallopeptidase, thrombospondin Adamts19 240322 2.04 type 1, 19 Wdr91 101240 WD repeat domain 91 2.01

- 176- Appendix *p value smaller than 0.05

Translation Cugbp2 14007 CUG triplet repeat, RNA binding protein 2 4.42* Eif1a 13664 Eukaryotic translation initiation factor 1A 4.16 Brunol4 108013 Bruno-like 4, RNA binding protein 3.52 Agxt2l1 71760 Alanine-glyoxylate aminotransferase 2-like 1 3.14 Gtpbp3 70359 GTP binding protein 3 3.08 Eif2ak1 15467 Eukaryotic translation initiation factor 2 α-kinase 1 3.05 Mrpl15 27395 Mitochondrial ribosomal protein L15 3.00 Trit1 66966 tRNA isopentenyltransferase 1 2.85 Rpl9 20005 Ribosomal protein L9 2.85 Asparaginyl-tRNA synthetase 2 mitochondrial Nars2 244141 2.62 (putative) Rps17 20068 Ribosomal protein S17 2.57 Trim6 94088 Tripartite motif-containing 6 2.39 Pcmtd1 319263 Protein-L-isoaspartate O-methyltransferase domain 1 2.31 Tnrc6b 213988 Trinucleotide repeat containing 6b 2.13 Heparan sulfate (glucosamine) 3-O-sulfotransferase Hs3st3b1 54710 2.08 3B1 Eef1g 67160 Eukaryotic translation elongation factor 1 gamma 2.07 *p value smaller than 0.05

G-protein signaling Fshr 14309 Follicle stimulating hormone receptor 4.63* Uts2r 217369 Urotensin 2 receptor 4.47 Gpr125 70693 G protein-coupled receptor 125 4.20 Csprs 114564 Component of Sp100-rs 4.16 Rgs9 19739 Regulator of G-protein signaling 9 2.61 Ptger3 19218 Prostaglandin E receptor 3 (subtype EP3) 2.44 Gpr50 14765 G-protein-coupled receptor 50 2.28 Guanine nucleotide binding protein, alpha Gnat1 14685 2.10 transducing 1 *p value smaller than 0.05

Cell adhesion Muc10 17830 Mucin 10, submandibular gland salivary mucin 6.40* Zp3r 22789 Zona pellucida 3 receptor 4.04 Cgn 70737 Cingulin 3.81 Sned1 208777 Sushi, nidogen and EGF-like domains 1 3.47 Mfap4 76293 Microfibrillar-associated protein 4 3.40 Gjd2** 14617 Gap junction protein, delta 2 3.37 Cd84 12523 CD84 antigen 3.33

- 177- Appendix

Alcam 11658 Activated leukocyte cell adhesion molecule 3.20 Vcan 13003 Versican 3.07 Fermt2 218952 Fermitin family homolog 2 2.86 Pvrl4 71740 Poliovirus receptor-related 4 2.77 Muc4 140474 Mucin 4 2.75 Cldn2 12738 Claudin 2 2.69 Ceacam1 26365 CEA-related cell adhesion molecule 1 2.55 Nrxn3 18191 Neurexin III 2.51 Ncam1 17967 Neural cell adhesion molecule 1 2.50 Limk2 16886 LIM motif-containing protein kinase 2 2.49 Itgb2l 16415 Integrin beta 2-like 2.48 Nfasc 269116 Neurofascin 2.47 Ephb2 13844 Eph receptor B2 2.41 Ncan 13004 Neurocan 2.37 Muc13 17063 Mucin 13, epithelial transmembrane 2.35 Gpa33 59290 Glycoprotein A33 (transmembrane) 2.27 Mog 17441 Myelin oligodendrocyte glycoprotein 2.10 Jam3 83964 Junction adhesion molecule 3 2.06 *p value smaller than 0.05 **Genes reportedly associated with bone metabolism

Chemotaxis Prok1 246691 Prokineticin 1 2.62* Hebp1 15199 Heme binding protein 1 2.13 Edn3 13616 Endothelin 3 2.01 *p value smaller than 0.05

Transport Gabrp 216643 Gamma-aminobutyric acid (GABA-A) receptor, pi 6.55* Smok3a 545814 Sperm motility kinase 3A 5.30 Rph3al 380714 Rabphilin 3A-like (without C2 domains) 4.57 Potassium voltage gated channel, Shab-rel. subfam., Kcnb2 98741 4.51 2 Cnnm3 94218 Cyclin M3 4.33 Clca3 23844 Chloride channel calcium activated 3 4.00 Slc15a1 56643 Solute carrier family 15 (oligopeptide transporter), 1 3.96 Calcium channel, voltage-dependent, α- 2/Δ-subunit Cacna2d2 56808 3.93 2 Rbp3 19661 Retinol binding protein 3, interstitial 3.83 Kcns2 16539 K+ voltage-gated channel, subfamily S, 2 3.78 Grin2d 14814 Glutamate receptor, ionotropic, NMDA2D, epsilon 4 3.70 Translocase, inner mitochondrial membr. 8 homolog Timm8a2 223262 3.58 a2

- 178- Appendix

Ryr2 20191 Ryanodine receptor 2, cardiac 3.49 Klc1 16593 Kinesin light chain 1 3.40 Potassium inwardly-rectifying channel, subfamily J, Kcnj9 16524 3.37 9 Ap2a2 11772 Adaptor protein complex AP-2, alpha 2 subunit 3.33 SH3-domain GRB2-like (endophilin) interacting Sgip1 73094 3.33 protein 1 Aqp4 11829 Aquaporin 4 3.29 Rtp3 235636 Receptor transporter protein 3 3.27 Slc35f1 215085 Solute carrier family 35, member F1 3.27 Transient receptor potential cation channel, family Trpm1 17364 3.26 M, 1 Aqp11 66333 Aquaporin 11 3.17 Rab3d 19340 RAB3D, member RAS oncogene family 3.03 Itpr2 16439 Inositol 1,4,5-triphosphate receptor 2 2.93 Transient receptor potential cation channel, family C, Trpc7 26946 2.92 7 Solute carrier organic anion transport family, member Slco6b1 67854 2.89 6b1 Transient receptor potential cation channel, Trpc5 22067 2.87 subfamily C, 5 Laptm4b 114128 Lysosomal-associated protein transmembrane 4B 2.84 Hbb-bh1 15132 Hemoglobin Z, beta-like embryonic chain 2.84 Ap3b2 11775 Adaptor-related protein complex 3, beta 2 subunit 2.83 Slc26a10 216441 Solute carrier family 26, member 10 2.70 Solute carrier family 7, cationic amino acid Slc7a13 74087 2.67 transporter, 13 Scn8a 20273 Sodium channel, voltage-gated, type VIII, alpha 2.61 Atp8b1 54670 ATPase, class I, type 8B, member 1 2.59 Csn1s2a 12993 Casein alpha s2-like A 2.58 Spnb3 20743 Spectrin beta 3 2.56 Ipo7 233726 Importin 7 2.56 Slc25a2 83885 Solute carrier family 25, member 2 2.56 Slc10a1 20493 Solute carrier family 10, member 1 2.54 Slc25a40 319653 Solute carrier family 25, member 40 2.52 Stam2 56324 Signal transducing adaptor molecule 2 2.51 Pts 19286 6-pyruvoyl-tetrahydropterin synthase 2.45 Ift172 67661 Intraflagellar transport 172 homolog 2.44 Stxbp5l 207227 Syntaxin binding protein 5-like 2.44 Sri 109552 Sorcin 2.43 Golt1b 66964 Golgi transport 1 homolog B 2.40 Jakmip1 76071 Janus kinase and microtubule interacting protein 1 2.33 Ipo8 320727 Importin 8 2.30 Ap3m2 64933 Adaptor-related protein complex 3, mu 2 subunit 2.30

- 179- Appendix

Pkd1l3 244646 Polycystic kidney disease 1 like 3 2.21 Slc25a25 227731 Solute carrier family 25, member 25 2.16 Solute carrier family 17 (sodium phosphate), member Slc17a2 218103 2.15 2 Pacsin3 80708 Protein kinase C, casein kinase substrate in neurons 3 2.12 D2hgdh 98314 D-2-hydroxyglutarate dehydrogenase 2.12 Ces2 234671 Carboxylesterase 2 2.12 Igh-6 16019 Immunoglobulin heavy chain 6 (heavy chain of IgM) 2.11 Slc5a9 230612 Solute carrier family 5, member 9 2.10 Actr1a 54130 ARP1 actin-related protein 1 homolog A (yeast) 2.10 Mfsd11 69900 Major facilitator superfamily domain containing 11 2.09 Best2 212989 Bestrophin 2 2.08 Sgpp2 433323 Sphingosine-1-phosphate phosphotase 2 2.08 Slc14a2 27411 Solute carrier family 14 (urea transporter), member 2 2.06 Kif13a 16553 Kinesin family member 13A 2.05 Slc24a2 76376 Solute carrier family 24, member 2 2.02 Atp1a1 11928 ATPase, Na+/K+ transporting, alpha 1 polypeptide 2.02 Transient receptor potential cation channel, Trpc4ap 56407 2.00 subfamily C, 4 Dnajc6 72685 DnaJ (Hsp40) homolog, subfamily C, member 6 2.00 Vps33a 77573 Vacuolar protein sorting 33A 2.00 *p value smaller than 0.05

Energy metabolism Cytochrome P450, family 4, subfamily a, polypeptide Cyp4a10 13117 6.22* 10 Man1a 17155 Mannosidase 1, alpha 4.34 Gsta2 14858 Glutathione S-transferase, alpha 2 (Yc2) 4.07 Gpd2 14571 Glycerol phosphate dehydrogenase 2, mitochondrial 3.87 Pcyt1b 236899 Phosphate cytidylyltransferase 1 choline, β-isoform 3.78 Large 16795 Like-glycosyltransferase 3.77 Akr1d1 208665 Aldo-keto reductase family 1, member D1 3.73 A 50518 Nonagouti 3.49 Tbc1d5 72238 TBC1 domain family, member 5 3.40 Cytochrome P450, family 2, subfamily a, polypeptide Cyp2a5 13087 3.24 5 Cytochrome P450, family 4, subfamily a, polypeptide Cyp4a31 666168 3.22 31 Snca 20617 Synuclein, alpha 3.09 Pank1 75735 Pantothenate kinase 1 3.01 Alg6 320438 Asparagine-linked glycosylation 6 homolog 2.98 Idi2 320581 Isopentenyl-diphosphate delta isomerase 2 2.83 Cytochrome P450, family 19, subfamily a, Cyp19a1 13075 2.78 polypeptide 1

- 180- Appendix

Mettl7a1 70152 Methyltransferase like 7A1 2.62 Ckmt1 12716 Creatine kinase, mitochondrial 1, ubiquitous 2.60 Pigc 67292 Phosphatidylinositol glycan anchor biosynthesis, C 2.57 Acad9 229211 Acyl-coenzyme A dehydrogenase family member 9 2.57 Pgk2 18663 Phosphoglycerate kinase 2 2.56 Ugt3a1 105887 UDP glycosyltransferases 3 family, polypeptide A1 2.56 Fggy 75578 FGGY carbohydrate kinase domain containing 2.54 Galntl4 233733 UDP-N-acetyl-alpha-D-galactosamine 4 2.53 Low density lipoprotein receptor-related associated Lrpap1 16976 2.53 protein 1 Cytochrome P450, family 26, subfamily a, Cyp26a1 13082 2.49 polypeptide 1 Cytochrome P450, family 2, subfamily j, polypeptide Cyp2j5 13109 2.48 5 Beta-GlcNAc beta 1,4-galactosyltransferase, B4galt6 56386 2.48 polypeptide 6 Asah3l 230379 N-acylsphingosine amidohydrolase 3-like 2.44 B4galnt2 14422 Beta-1,4-N-acetyl-galactosaminyl transferase 2 2.42 Hydroxy-delta-5-steroid dehydrogenase, 3 beta- Hsd3b6 15497 2.42 isomerase 6 Alms1 236266 Alstrom syndrome 1 homolog 2.38 Bdh2 69772 3-hydroxybutyrate dehydrogenase, type 2 2.33 Indol1 209176 Indoleamine-pyrrole 2,3 dioxygenase-like 1 2.26 Protein kinase, AMP-activated, gamma 2 non- Prkag2 108099 2.23 catalytic subunit Cryzl1 66609 Crystallin, zeta (quinone reductase)-like 1 2.20 Apoa5 66113 Apolipoprotein A-V 2.20 Man2a2 140481 Mannosidase 2, alpha 2 2.17 Nagk 56174 N-acetylglucosamine kinase 2.16 Gstm3 14864 Glutathione S-transferase, mu 3 2.13 Gldc 104174 Glycine decarboxylase 2.11 Man1a2 17156 Mannosidase, alpha, class 1A, member 2 2.07 Phenazine biosynthesis-like protein domain Pbld 68371 2.06 containing Ddah1 69219 Dimethylarginine dimethylaminohydrolase 1 2.04 Neu3 50877 Neuraminidase 3 2.04 Bhmt 12116 Betaine-homocysteine methyltransferase 2.02 N-acylsphingosine amidohydrolase (acid Asahl 67111 2.00 ceramidase)-like Beta-GlcNAc beta 1,3-galactosyltransferase, B3galt5 93961 2.00 polypeptide 5 Aldh1l1 107747 Aldehyde dehydrogenase 1 family, member L1 2.00 *p value smaller than 0.05

Neurophysiological process

- 181- Appendix

Syn2 20965 Synapsin II 4.16* Gria4 14802 Glutamate receptor, ionotropic, AMPA4 (alpha 4) 3.52 Slitrk5 75409 SLIT and NTRK-like family, member 5 3.27 Grid2 14804 Glutamate receptor, ionotropic, delta 2 3.26 Grik1 14805 Glutamate receptor, ionotropic, kainate 1 2.66 Gria2 14800 Glutamate receptor, ionotropic, AMPA2 (alpha 2) 2.42 Chgb 12653 Chromogranin B 2.35 Syn3 27204 Synapsin III 2.13 *p value smaller than 0.05

Unknown process Rshl2a 66832 Radial spokehead-like 2A 5.34* Rex2 19715 Reduced expression 2 5.09 Apol7b 278679 Apolipoprotein L 7b 4.62 C77717 97361 Expressed sequence C77717 4.58 C79741 97877 Expressed sequence C79741 4.46 Tmem207 224058 Transmembrane protein 207 4.36 DXErtd11e 52003 DNA segment, Chr X, ERATO Doi 11, expressed 4.21 D15Ertd50e 52193 DNA segment, Chr 15, ERATO Doi 50, expressed 4.14 Lce1g 66195 Late cornified envelope 1G 4.02 D9Ertd496e 52361 DNA segment, Chr 9, ERATO Doi 496, expressed 3.85 D5Ertd163e 52181 DNA segment, Chr 5, ERATO Doi 163, expressed 3.84 Tex13 83555 Testis expressed gene 13 3.66 D3Ertd246e 52211 DNA segment, Chr 3, ERATO Doi 246, expressed 3.55 D8Ertd620e 52423 DNA segment, Chr 8, ERATO Doi 620, expressed 3.45 D2Ertd239e 51880 DNA segment, Chr 2, ERATO Doi 239, expressed 3.42 D4Ertd571e 52341 DNA segment, Chr 4, ERATO Doi 571, expressed 3.22 Lrrc52 240899 Leucine rich repeat containing 52 3.06 D9Ertd26e 52062 DNA segment, Chr 9, ERATO Doi 26, expressed 3.04 Prr18 320111 Proline rich region 18 3.00 Pramel5 384077 Preferentially expressed antigen in melanoma like 5 2.91 Morn1 76866 MORN repeat containing 1 2.82 D5Ertd521e 52350 DNA segment, Chr 5, ERATO Doi 521, expressed 2.79 Fbxw14 50757 F-box and WD-40 domain protein 14 2.77 Oas1d 100535 2'-5' oligoadenylate synthetase 1D 2.74 D3Ertd162e 52252 DNA segment, Chr 3, ERATO Doi 162, expressed 2.72 Oas1e 231699 2'-5' oligoadenylate synthetase 1E 2.65 D2Ertd105e 52165 DNA segment, Chr 2, ERATO Doi 105, expressed 2.63 Ankrd34a 545554 Ankyrin repeat domain 34A 2.51 Phxr5 18690 Per-hexamer repeat gene 5 2.51 Tmem109 68539 Transmembrane protein 109 2.49 D5Ertd615e 52401 DNA segment, Chr 5, ERATO Doi 615, expressed 2.44

- 182- Appendix

Lce1i 76585 Late cornified envelope 1I 2.37 D7Ertd1e 51959 DNA segment, Chr 7, ERATO Doi 1, expressed 2.35 D17Ertd657e 52029 DNA segment, Chr 17, ERATO Doi 657, expressed 2.33 Ankrd41 234396 Ankyrin repeat domain 41 2.32 Rtbdn 234542 Retbindin 2.31 Akr1c20 116852 Aldo-keto reductase family 1, member C20 2.29 Wdr20b 70948 WD repeat domain 20b 2.28 D7Ertd59e 52006 DNA segment, Chr 7, ERATO Doi 59, expressed 2.27 D15Ertd180e 52498 DNA segment, Chr 15, ERATO Doi 180, expressed 2.26 D18Ertd653e 52662 DNA segment, Chr 18, ERATO Doi 653, expressed 2.25 Tmem86b 68255 Transmembrane protein 86B 2.19 BC048546 232400 cDNA sequence BC048546 2.18 D2Ertd127e 51873 DNA segment, Chr 2, ERATO Doi 127, expressed 2.16 Krt2-ps1 64819 Keratin complex 2, basic, pseudogene 1 2.14 Heatr5b 320473 HEAT repeat containing 5B 2.14 D5Ertd215e 52224 DNA segment, Chr 5, ERATO Doi 215, expressed 2.14 Dcpp3 620253 Demilune cell and parotid protein 3 2.08 BC050210 381337 cDNA sequence BC050210 2.07 BC016548 211039 cDNA sequence BC016548 2.04 BC031353 235493 cDNA sequence BC031353 2.02 BC030500 234290 cDNA sequence BC030500 2.02 *p value smaller than 0.05

- 183- Appendix Table A5. List of load-regulated molecular pathways in single loading

Molecular Pathway Counts Map ID Translation _Regulation of translation initiation 60.00%* 498 Cytoskeleton remodeling_α-1A adrenergic receptor-dependent 66.67% 2395 inhibition of PI3K Immune response _IL22 signaling pathway 50.00% 522 Immune response _Antigen presentation by MHC class I 42.31% 2100 Immune response _PIP3 signaling in B lymphocytes 38.71% 702 Immune response _CD28 signaling 35.14% 620 Immune response _BCR pathway 33.33% 655 Signal transduction_Activation of PKC via G-Protein coupled receptor 33.33% 453 Development_Transcription regulation of granulocyte development 35.48% 458 Signal transduction_PKA signaling 39.13% 675 Immune response _ICOS-ICOSL pathway in T-helper cell 33.33% 619 Immune response _NFAT in immune response 33.33% 668 Development_VEGF signaling and activation 34.38% 539 Signal transduction_Calcium signaling 34.38% 550 Cell cycle_Spindle assembly and chromosome separation 34.38% 712 Apoptosis and survival_Anti-apoptotic action of membrane-bound 37.50% 2736 ESR1 Cardiac Hypertrophy_Ca(2+)-dependent NF-AT signaling in Cardiac 32.43% 2234 Hypertrophy Immune response_Inhibitory action of Lipoxins on TNF-alpha 32.35% 2727 signaling Transcription_Transcription factor Tubby signaling pathways 50.00% 461 Apoptosis and survival_Inhibition of ROS-induced apoptosis by 34.62% 2740 17beta-estradiol Development_EPO-induced Jak-STAT pathway 31.43% 737 Muscle contraction_ACM regulation of smooth muscle contraction 31.43% 2657 Oxidative stress_Role of ASK1 under oxidative stress 36.36% 521 Immune response _CXCR4 signaling via second messenger 36.36% 613 Signal transduction_IP3 signaling 30.00% 557 Regulation of CFTR activity (norm and CF) 30.00% 2269 Cell cycle_Nucleocytoplasmic transport of CDK/Cyclins 42.86% 473 Apoptosis and survival_Anti-apoptotic TNFs/NF-kB/IAP pathway 33.33% 721 Apoptosis and survival_Anti-apoptotic TNFs/NF-kB/Bcl-2 pathway 29.27% 720 Transcription_Ligand-Dependent Transcription of Retinoid-Target 31.25% 369 genes Transport_RAN regulation pathway 40.00% 404 Immune response _PGE2 in immune and neuroendocrine system 33.33% 2387 interactions TCA (tricarboxylic acid cycle) 35.00% 814 Cytoskeleton remodeling_Neurofilaments 31.03% 1491 Transcription_Ligand-dependent activation of the ESR1/SP pathway 31.03% 2208 Development_TGF-beta receptor signaling 29.41% 475 Immune response_Bacterial infections in normal airways 28.21% 2694 Development_IGF-RI signaling 27.27% 540

- 184- Appendix Transport_ACM3 in salivary glands 32.00% 2652 Immune response _PGE2 common pathways 30.00% 2386 Immune response _CCR3 signaling in eosinophils 25.45% 736 Chemotaxis_Lipoxin inhibitory action on neutrophil migration 28.57% 2731 Development_A3 receptor signaling 28.57% 644 Transcription_CREB pathway 28.57% 409 Development_PIP3 signaling in cardiac myocytes 27.50% 701 Cell cycle_Regulation of G1/S transition (part 2) 30.77% 474 Development_Mu-type opioid receptor signaling 30.77% 2453 Immune response _IL10 signaling pathway 35.29% 531 DNA damage_Role of SUMO in p53 regulation 35.29% 648 Neurophysiological process_Mu-type opioid receptor-mediated 35.29% 2450 analgesia Development_TPO signaling via JAK-STAT pathway 31.82% 469 Immune response _MIF - the neuroendocrine-macrophage connector 29.63% 518 Cell cycle_Role of APC in cell cycle regulation 28.13% 472 Immune response _Fc epsilon RI pathway 26.19% 566 Signal transduction_AKT signaling 27.03% 554 Development_A2A receptor signaling 27.03% 643 Inhibitory action of Lipoxins on neutrophil migration 27.03% 2692 Development_EGFR signaling via PIP3 33.33% 692 G-protein signaling_G-Protein alpha-s signaling cascades 30.43% 640 CFTR folding and maturation (norm and CF) 35.71% 2688 Transport_Alpha-2 adrenergic receptor regulation of ion channels 28.57% 2432 Mucin expression in CF via IL-6, IL-17 signaling pathways 28.57% 2655 Cell cycle_Regulation of G1/S transition (part 1) 26.32% 544 Immune response _IL1 signaling pathway 27.27% 658 Signal transduction_cAMP signaling 27.27% 660 *p value smaller than 0.05

- 185- Appendix Table A6. List of load-regulated molecular pathways in repeated loading Map Molecular Pathway Counts ID Cell adhesion_Endothelial cell contacts by non-junctional mechanisms 79.17%* 727 Cell adhesion_Integrin-mediated cell adhesion and migration 62.22% 450 Cytoskeleton remodeling_Integrin outside-in signaling 60.87% 664 Cytoskeleton remodeling_FAK signaling 59.57% 449 Signal transduction_IP3 signaling 63.16% 557 Translation _Regulation activity of EIF2 63.89% 497 Development_Endothelin-1/EDNRA signaling 60.98% 2255 Cytoskeleton remodeling_Cytoskeleton remodeling 48.96% 714 Cell adhesion_Role of tetraspanins in the integrin-mediated cell adhesion 62.16% 2023 Transcription_Ligand-Dependent Transcription of Retinoid-Target genes 62.50% 369 Transcription_CREB pathway 60.00% 409 G-protein signaling_Proinsulin C-peptide signaling 57.89% 2815 Signal transduction_Erk Interactions: Inhibition of Erk 61.29% 447 Cell adhesion_Chemokines and adhesion 46.24% 716 Apoptosis and survival_BAD phosphorylation 57.14% 661 Development_Alpha-2 adrenergic receptor activation of ERK 55.00% 2427 Development_EDG3 signaling pathway 61.54% 2951 Development_ACM1, ACM3, ACM5 activation of ERK 55.56% 2518 Development_A3 receptor signaling 55.56% 644 Development_Flt3 signaling 53.66% 2236 Cardiac Hypertrophy_NF-AT signaling in Cardiac Hypertrophy 49.15% 2235 Immune response _CCR3 signaling in eosinophils 49.15% 736 Neurophysiological process_HTR1A receptor signaling in neuronal cells 58.62% 2946 Development_PDGF signaling via MAPK cascades 55.88% 654 Development_EDG1 signaling pathway 59.26% 2809 Development_VEGF-family signaling 56.25% 445 Signal transduction_Calcium signaling 56.25% 550 Immune response_Function of MEF2 in T lymphocytes 52.38% 541 Cell adhesion_Histamine H1 receptor signaling 54.05% 2435 Cytoskeleton remodeling_α-1A adrenergic receptor-dependent inhibition 75.00% 2395 of PI3K Regulation of lipid metabolism_Insulin regulation of glycogen 52.50% 725 metabolism Immune response _Role of the C5b-9 complement complex in cell 60.87% 459 survival Immune response_HTR2A-induced activation of cPLA2 57.14% 2439 Cytoskeleton remodeling_Fibronectin-binding integrins in cell motility 57.14% 451 Development_FGFR signaling pathway 51.16% 444 Immune response_CD28 signaling 51.16% 620 Development_Dopamine D2 receptor transactivation of PDGFR in non- 61.90% 2456 neuronal cells Development_A2B receptor: action via G-protein alpha s 52.63% 482 Translation_Insulin regulation of translation 52.63% 723 Development_Ligand-independent activation of ESR1 and ESR2 52.63% 2210 Development_ACM2 and ACM4 activation of ERK 54.55% 2516

- 186- Appendix Cytoskeleton remodeling_Role of PKA in cytoskeleton reorganisation 54.84% 543 Development_Role of CDK5 in neuronal development 54.84% 2368 Cell adhesion_Alpha-4 integrins in cell migration and adhesion 54.84% 2749 Neurophysiological process_Receptor-mediated axon growth repulsion 50.00% 527 Cell adhesion_Integrin inside-out signaling 50.00% 710 Development_VEGF signaling via VEGFR2 - generic cascades 55.17% 533 Neurodisease_Parkin disorder under Parkinson's disease 55.17% 666 Neurophysiological process_ACM regulation of nerve impulse 52.94% 2656 Development_Neurotrophin family signaling 51.35% 636 Cardiac Hypertrophy_Ca(2+)-dependent NF-AT signaling in Cardiac 51.35% 2234 Hypertrophy Development_MAG-dependent inhibition of neurite outgrowth 55.56% 735 Development_Activation of ERK by Alpha-1 adrenergic receptors 53.13% 2392 Development_Dopamine D2 receptor transactivation of EGFR 60.00% 2457 Regulation of CFTR activity (norm and CF) 50.00% 2269 Immune response_NTS activation of IL-8 in colonocytes 51.43% 722 Development_Angiotensin activation of Akt 56.00% 436 G-protein signaling_G-Protein alpha-q signaling cascades 56.00% 639 Development_EGFR signaling via PIP3 61.11% 692 Neurophysiological process_Dopamine D2 receptor transactivation of 61.11% 2455 PDGFR in CNS Muscle contraction_ GPCRs in the regulation of smooth muscle tone 46.30% 2393 Transcription_PPAR Pathway 50.00% 546 Immune response_Fc epsilon RI pathway 47.83% 566 Transport_Clathrin-coated vesicle cycle 44.12% 2640 Development_Glucocorticoid receptor signaling 56.52% 410 Transport_Alpha-2 adrenergic receptor regulation of ion channels 53.57% 2432 Neurophysiological process_EphB receptors in dendritic spine 53.57% 528 morphogenesis Membrane-bound ESR1: interaction with G-proteins signaling 50.00% 2212 Development_G-Proteins mediated regulation MARK-ERK signaling 50.00% 463 Immune response_MIF - the neuroendocrine-macrophage connector 51.62% 518 Cell cycle_Regulation of G1/S transition (part 2) 53.85% 474 Cell adhesion_Endothelial cell contacts by junctional mechanisms 53.85% 745 Cytoskeleton remodeling_TGF, WNT and cytoskeletal remodeling 40.19% 715 G-protein signaling_Regulation of p38 and JNK signaling mediated by 50.00% 455 G-proteins Transcription_Transcription factor Tubby signaling pathways 66.67% 461 WtCFTR and delta508 traffic / Clathrin coated vesicles formation (norm 57.89% 2672 and CF) Development_EDG1 signaling via beta-arrestin 51.72% 2808 Development_A2A receptor signaling 48.65% 643 Regulation of lipid metabolism_ACM stimulation of Arachidonic acid 45.83% 2658 production Blood coagulation_GPCRs in platelet aggregation 44.44% 2442 Immune response_MIF-JAB1 signaling 54.55% 520 Oxidative stress_Role of ASK1 under oxidative stress 54.55% 521 Development_Kappa-type opioid receptor activation of ERK 54.55% 2552 Development_Endothelin-1/EDNRA transactivation of EGFR 48.57% 2254

- 187- Appendix Neurophysiological process_Glutamate regulation of Dopamine D1A 50.00% 2425 receptor Apoptosis and survival_HTR1A signaling 47.37% 2947 Translation _Regulation of translation initiation 52.00% 498 Development_Delta-type opioid receptor mediated cardioprotection 52.00% 2666 Regulation of lipid metabolism_Insulin signaling:generic cascades 46.34% 724 Development_GDNF signaling 55.00% 646 Development_Delta-type opioid receptor signaling via G-protein alpha- 55.00% 2664 14 G-protein signaling_G-Protein alpha-12 signaling pathway 48.48% 454 Signal transduction_Activation of PKC via G-Protein coupled receptor 45.45% 453 Development_Angiotensin signaling via PYK2 47.22% 438 Development_EGFR signaling pathway 44.00% 443 Cytoskeleton remodeling_Regulation of actin cytoskeleton by Rho 52.17% 551 GTPases G-protein signaling_G-Protein beta/gamma signaling cascades 52.17% 641 Regulation of lipid metabolism_α-1 adrenergic receptors signaling 46.15% 2385 Transcription_Assembly of RNA Polymerase II preinitiation complex 55.56% 673 Signal transduction_PTEN pathway 45.24% 676 Development_Membrane-bound ESR1: interaction with growth factors 47.06% 2211 signaling Development_EDNRB signaling 47.06% 2273 Development_Angiotensin activation of ERK 50.00% 437 Transcription_Role of Akt in hypoxia induced HIF1 activation 50.00% 448 Cell adhesion_Cadherin-mediated cell adhesion 50.00% 2122 Cytoskeleton remodeling_ACM3 and ACM4 in keratinocyte migration 52.38% 2651 G-protein signaling_EDG5 signaling 52.38% 2814 Development_EGFR signaling via small GTPases 48.28% 704 Cell cycle_Role of Nek in cell cycle regulation 48.28% 731 Development_GDNF family signaling 45.00% 495 Translation_Translation regulation by Alpha-1 adrenergic receptors 45.00% 2390 Cell cycle_Spindle assembly and chromosome separation 46.88% 712 Proteolysis_Role of Parkin in the Ubiquitin-Proteasomal Pathway 50.00% 662 Apoptosis and survival_Anti-apoptotic action of membrane-bound ESR1 50.00% 2736 Regulation of lipid metabolism_Insulin regulation of fatty acid 43.48% 726 methabolism Transcription_Receptor-mediated HIF regulation 45.71% 416 Muscle contraction_ACM regulation of smooth muscle contraction 45.71% 2657 Development_EDG6 signaling pathway 57.14% 2952 Transport_RAB3 regulation pathway 57.14% 406 Cytoskeleton remodeling_Reverse signaling by ephrin B 47.67% 529 Cytoskeleton remodeling_Slit-Robo signaling 47.67% 2121 Immune response_PGE2 common pathways 47.67% 2386 Cytoskeleton remodeling_Neurofilaments 48.00% 1491 Immune response_BCR pathway 42.86% 655 Immune response_NFAT in immune response 42.86% 668 Transcription_Androgen Receptor nuclear signaling 42.86% 2202 G-protein signaling_G-Protein alpha-s signaling cascades 46.43% 640 Immune response _IFN gamma signaling pathway 42.22% 432

- 188- Appendix G-protein signaling_Cross-talk between Ras-family GTPases 50.00% 408 Development_WNT signaling pathway. Part 1 50.00% 515 Oxidative phosphorylation 37.37% 920 Cell adhesion_ECM remodeling 41.18% 717 Apoptosis and survival_Role of CDK5 in neuronal death and survival 45.16% 2374 dATP/dITP metabolism 40.74% 865 Anandamide biosynthesis and metabolism 53.33% 2428 Muscle contraction_Delta-type opioid receptor in smooth muscle 53.33% 2663 contraction Development_VEGF signaling and activation 44.12% 539 Development_Leptin signaling via PI3K-dependent pathway 44.12% 719 Development_Regulation of CDK5 in CNS 47.83% 2226 Development_Mu-type opioid receptor regulation of proliferation 47.83% 2424 Immune response _ICOS pathway in T-helper cell 43.24% 619 NGF activation of NF-kB 46.15% 653 Immune response_Antigen presentation by MHC class I 46.15% 2100 Immune response_IL-2 activation and signaling pathway 41.30% 430 ATP/ITP metabolism 37.65% 873 Translation _Regulation activity of EIF4F 40.82% 496 Development_WNT signaling pathway. Part 2 40.82% 516 Transport_RAN regulation pathway 50.00% 404 Neurophysiological process_Delta-type opioid receptor in nervous 50.00% 2567 system Cell cycle_Role of SCF complex in cell cycle regulation 44.83% 706 Development_HGF signaling pathway 43.75% 530 Development_Beta-adrenergic receptors transactivation of EGFR 43.75% 2433 Transcription_ChREBP regulation pathway 53.85% 464 *p value smaller than 0.05

- 189- Curriculum Vitae

Curriculum Vitae

Elad Wasserman

Born 27th September 1972, Kiev, Ukraine

2004 - 2010 Ph.D. Thesis “Differentially load-regulated genes in mouse trabecular osteocytes” under the supervision of Prof. Dr. Ralph Müller and Prof. Dr. Itai Bab

2000 – 2003 QBI Inc., Nes-Ziona, Israel, Laboratory Technician

1998 - 2000 Tel-Aviv University, Sackler School of Medicine, Department of Cell Biology and Histology, Master of Science

1995 - 1997 Tel-Aviv University, Faculty of Life Sciences, Bachelor of Science

1991 - 1994 Ukrainian State Medical University, Faculty of Dentistry, Kiev, Ukraine

1988 - 1991 Ukrainian Medical College, Department of Nursing, Kiev, Ukraine

1980 - 1988 Primary School in Kiev, Ukraine

- 190-