Inflammatory Gene Induction in Meniscal Fibrochon- Drocytes

BIOMECHANICAL SIGNALS MEDIATE CELLULAR MECHANO- TRANSDUCTION AND GENE REGULATION

DISSERTATION

Presented in Partial Fulfillment for the Requirements for the Degree Doctor of Philosophy in the Graduate School of The Ohio State University

By Shashi Madhavan, D.D.S, M.D.S, M.S * * * * *

The Ohio State University 2007

Dissertation Committee: Approved by Dr. Sudha Agarwal, Adviser Dr. Doug Kniss Dr. David Padgett Adviser Dr. Sarandeep Huja Graduate Program in Dr. John Walters Oral Biology

ABSTRACT

Cartilage is a mechanosensitive tissue, which can perceive and respond to biomechnical signals. Despite the known importance of biomechanical signals in the etiopathogenesis of arthritic diseases, and their effectiveness in joint restoration, little is understood about their actions at the cellular level. Recent molecular approaches have revealed that specific biomechanical stimuli and cell interactions generate intracellular signals that are powerful inducers or suppressors of proinflammatory and reparative genes in chondrocytes. Bio- mechanical signals are perceived by cartilage in a magnitude, frequency, and time dependent manner. Static as well as dynamic biomechanical forces of high magnitudes induce proinflammatory genes and inhibit matrix synthesis. Contrarily, dynamic biomechanical signals of low / physiological magnitudes are potent anti-inflammatory signals that inhibit IL-1β-induced proinflammatory gene transcription, as well as abrogate IL-1β

/ TNF-α-induced inhibition of matrix synthesis. Recent studies have identified NF-κB transcription factors as key regulators of biomechanical signals-mediated proinflammatory as well as anti-inflammatory actions. These signals intercept multiple steps in the

NF-κB signaling cascade to regulate cytokine gene expression. Taken together these findings provide insight into how biomechanical signals regulate inflammatory and reparative

gene transcription, underscoring their potential in enhancing the ability of chondrocytes to curb inflammation in diseased joints.

iii

DEDICATION

Dedicated to Amma

ACKNOWLEDGMENTS

At the outset I would like to extend my heartfelt gratitude and thanks to my adviser Dr.

Sudha Agarwal. I am indebted to her, because without her support and guidance, my achievements would not have been possible. I am also extremely thankful, to Dr. James

Deschner, Dr. Tom Knobloch and Dr. Mirela Anghelina for offering their valuable support and guidance through out the period of my study. I thank the members of my dissertation committee, Dr. Doug Kniss, Dr. Saran Huja, Dr. David Padgett and Dr. John Wal- ters, for graciously consenting to be part of my thesis committee and for giving me valuable inputs and ideas. The other members in my lab, Agata, Bessie, Bjoern, Danen, Jin,

Priyangi and Ravi, also need to be thanked for their support.

My family has been a big source of support for me in all my endeavors. My mother, who passed away on the first of July this year, was a big source of love, support and inspira- tion for me. Everything that I have achieved so far is because of her unflinching love and words of advice. I love you mom and your memories keep me going.

VITA

06.21.1976...... …..Born - Chennai, India

1995-2000 Ragas Dental College Bachelor of Dentistry University of Madras Dental Surgery TamilNadu, India.

2001-2004 Madras Dental College Master of Orthodontics University of Madras Dental Surgery TamilNadu, India.

2004-2007 The Ohio State University Master of Oral Biology Columbus, Ohio Science

2004-2007 The Ohio State University PhD Oral Biology Columbus, Ohio

PUBLICATIONS

1. Madhavan S, Anghelina M, Rath-Deschner B, Wypasek E, John A, Deschner J, Piesco N, Agarwal S. Biomechanical signals exert sustained attenuation of proinflammatory gene induction in articular chondrocytes. Osteoarthritis & Cartilage. May 25, 2006; [Epub ahead of print]

2. Ferretti M*, Madhavan S*, Deschner J, Rath-Deschner B, Wypasek E, Agarwal S. Dy- namic biophysical strain modulates proinflammatory gene induction in meniscal fibrochondrocytes. Am J Physiol Cell Physiol. 2006; 290(6):C1610-5. Epub Feb 1,2006. (* co-first authors)

3. Dossumbekova A, Anghelina M, Madhavan S, He L, Quan N, Knobloch T, Agarwal S. Inhibition of IKK activity mediates attenuation of proinflammatory gene induction by Biome- chanical signals in chondrocytes. Arthritis and Rheumatism, In press, 2007.

4. Knobloch T.J, Madhavan S, Nam J, Agarwal S (Jr), Agarwal S. Regulation of Chondro- cytic Gene Expression by Biomechanical Signals. Criticals Reviews in Eukaryotic Gene Ex- pression. In press, 2008

5. Madhavan S, Anghelina M , Sjostrom D, Dossumbekova A , Guttridge D, Agarwal S . Biomechanical Signals suppress TAK1 activation to Inhibit Proinflammatory Gene Induction. Journal of Immunology, November, 2007

6. Madhavan S, Liu Z, King G.J, Agarwal S. Regulation of Bone remodeling by Mechanical forces. Journal of Bone and Mineral Research, 2007 (In review)

FIELDS OF STUDY

Major Field: Oral Biology vii

TABLE OF CONTENTS

Page

Abstract ...... ii

Dedication...... iv

Acknowledgments...... v

Vita ...... vi

List of figures...... ix

List of symbols...... xi

Chapters 1. Introduction………………………………………………………………………. 1 2. Dynamic Biophysical Strain Modulates Proinflammatory Gene Induction In Meniscal Fibrochondrocytes……………………………………………………..17 3. Biomechanical Signals Exert Sustained Attenuation of Proinflammatory Gene induction In Articular Chondrocytes………………………………………39 4. Biomechanical Signals suppress TAK1 activation to Inhibit Proinflammatory Gene Induction in Fibrochondrocytes………………………..70 5. Differential regulation of proinflammatory genes by mechanical signals leads to alveolar bone remodeling…………………………………………… .102 6. Conclusion……………………………………………………………………..130 Bibliography…………………………………………………………………...... 133

viii

LIST OF FIGURES

Figure Page

1. Regulation of IL-1β -dependent iNOS induction by dynamic tensile force

at various magnitudes and frequencies in meniscal fibrochondrcocytes………...33

2. Effect of DTF on IL-1β-dependent MMP-13 and TNF-α induction in

fibrochondrocytes………………………………………………………………..35

3. Temporal regulation of IL-1β-induced iNOS induction by dynamic

tensile strain……………………………………………………………………..37

4. Effect of DTF on iNOS induction in articular chondrocytes…………………….62

5. Effect of DTF on IL-1β-dependent Cox-2 induction in articular chondrocytes…64

6. DTF-induced blocking of IL-1β-dependent induction of MMP-9 and MMP-13..66

7. Effect of DTF on IL-1β induced attenuation of aggrecan synthesis……………..68

8. CTS inhibits IL-1β, TNF-α, and LPS induced proinflammatory mRNA

expression……………………………………………………………..…………89

9. CTS suppresses IL-1β-induced TAK1 activation and IκBα activation……...... 91

10. CTS inhibits IL-1β-induced I-κBβ degradation…………………………………94

11. CTS inhibits IL-1β-induced NF-κB phosphorylation, nuclear translocation and

DNA binding…………………………………………………………………….96

12. Schematic representation of the mechanisms of intracellular actions of

CTS.…………………………………………………………………………….100

13. LTF inhibits synthesis of pro-inflammatory mediators COX-2 and IL-1β in

Rat PDL cells…………………………………………………………………...119

14. LTF and HTF differentially regulate COX-2 and IL-1β mRNA expression

in osteoblast like PDL cells……………………………………………………..122

15. LTF inhibits the nuclear translocation, mRNA expression, and protein synthesis

of NF-κB p65…………………………………………………………………..124

16. Regulation of I-κB degradation and synthesis by LTF…………………………126

17. LTF acts as an anabolic signal………………………………………………….128

LIST OF SYMBOLS

CTS – Cyclic tensile strain CCF – Cyclic compressive force DTF – Dynamic tensile forces BMP - bone morphogenic protein COX-2 - cyclooxygenase-2 EMSA - electrophoretic mobility shift ERK - extracellular signal related kinase FGF – fibroblast growth factor IKK - I-kB kinase IKK-α - IkB kinase subunit alpha IKK-β - IkB kinase subunit beta IKK-γ- IkB kinase subunit gamma IFN - interferon IL-1R - interleukin 1 receptor IL-1β - interleukin 1β IL-6 - interleukin 6 iNOS - inducible nitric oxide synthase IRAK - IL-1R associated kinase I-κB - inhibitor of NF-κB MMP - matrix metalloprotease NF-κB - nuclear factor kappa B NIK - NF-kB inhibitor kinase NO - nitric oxide OA/RA – Osteo-, Rheumatoid arthritis xi

PCR - Polymerase chain reaction PGE2 - Prostaglandin E2 QCRT/PCR - quantitative PCR rhIL-1β - recombinant human IL-1β RT/PCR - reverse transcriptase/PCR TAB-1 – TAK binding protein-1 TAK-1 – TGF-β activating kinase TMJ - temporomandibular joint TMJD - TMJ disorder TGF-β - transforming growth factor β TIMP -tissue inhibitor metalloprotease TNF-α - tumor necrosis factor TRAF-6-TNFR associated factor -6

xii

CHAPTER 1

INTRODUCTION

Biomechanical signals play a major role in the development and homeostasis of the musculoskeletal system. Cytoskeletal rearrangments provide the morphogenic signals necessary for embryonic cell movement and force generation.

Morphogenesis is intrinsically a cascade of biomechanical signal transductions.

These signals direct the molecular events and genetic cues that promote embryonic pattern formation and tissue differentiation. Biomechanical signals provide the bridge between gross morphological signals and molecular gene expression. Compressive, tensile and shear forces interplay to orchestrate a complex and tightly regulated series of events that ultimately direct gene expression within the cell. While nearly all cell types examined demonstrate significant responses to biomechanical stimuli (osteoblasts, chondrocytes, myob- lasts, mesechymal, epithelial, endothelial, and fibroblastic cells), the precise patterns of gene expression that manifest are less well characterized. The functional responses to biophysical forces, on the other hand, have been well documented in the literature (1).

This dissertation focuses on some of the recent stimulating findings centered on the specific transcriptional responses of cartilage and chondrocytes to external biomechanical stimuli. The link between mechanical signals, mechanotransduction, and gene expression is only recently being fully experimentally ex- plored. It is clear that biomechanical signal transduction progresses from the single cell intracellular environment, to apposed cells and extracellular matri- ces, to the tissue microenvironment, to the organ macroenvironment, and even- tually is functionally manifested at the organism level.

There are three types of cartilagenous tissues that perceive biomechanical stimuli: hyaline articular cartilage, fibrocartilage and elastic cartilage. Hyaline articular cartilage is an avascular tissue present at articulating surfaces that functions to absorb loads and dissipate the frictional forces realized at these joints.

Fibrocartilage and elastic cartilage are both vascularized. While elastic cartilage provides flexibility to the tissues, fibrocartilage takes part in absorbing loads and facilitating smooth joint movement.

Cartilage, Chondrocytes and Biomechanical Forces

Under physiological conditions, articular cartilage is often simultaneously exposed to axial compressive loading (normal to articular surface), lateral radial and circumferential tensile strain, and fluid shear forces. This complex model-

ing may be reproduced to varying degrees in vitro using cartilage explants, chondrocyte-seeded tissue engineered scaffoldings, and monolayered chondrocyte cell culture systems. The physiological recapitulation of the explant models makes them attractive systems for mimicking the in situ complexity of joint movement kinematics. However, the reductionistic approach of isolated chondrocytes in 3D scaffolding environments or monolayers on silastic stretchable membranes permits a simplified, tightly controlled and molecularly dissectable model system. It is clearly a combination of these different models that will ultimately allow investigators to delineate the complicated signaling pathways activated in response to physiological mechanical forces in the in situ tissue microenvironent.

The structure of hyaline (articular) cartilage allows the tissue to respond to mechanical forces in unique ways. The extracellular matrix (ECM) of cartilage can be considered a viscoelastic material with a rigid organic component of collagen and aggregated proteoglycans (PGs) integrated with a hydrated interstitial fluid component to present as a structured porous matrix. The combination of fluid-flow generated signals coupled to matrix mechanotransduction generates a complex series of signaling cascades, and ultimately, a biomechanical signal-dependent transcriptional response. It is estimated that only 5% of the volume of articular cartilage can be accounted for by the resident cell population of chondrocytes, leaving the vast majority of the cartilage volume occupied by the ECM and interstitial fluids. It is this network of fibrillar colla-

gen, especially collagen type II and negatively charged aggrecans that defines the signal transduction matrix essential for chondrocyte stimulation. This charged fluidi matrix of crosslinked proteoglycans and electrolytes permits the electrostatic osmotic pressure regulation of the dynamic stiffness, as well as re- silience to frictional forces, of the articular cartilage under physiological conditions (2).

The biomechanical signals generated during joint movement are essential components of the cells’ and tissue’s abilities to repair and recover following physiological insults, as well as maintain homeostasis. Chondrocytes are positioned as primary responders to biomechanical signals as they are transduced through the viscoelastic ECM. It is well documented that chondrocytes in cartilage are mechanosensitive cells and can perceive mechanical signals and respond to them by converting physical signals into biochemical events.

Compressive Forces Regulate Cartilage Damage and Repair

The normal physiological movement of joints exerts compressive loading on articular cartilage with either intermittent or cyclic durations. This dynamic compression of the ECM results in synovial fluid exudation and re-absorption that facilitates the diffusion of nutrients and oxygen into avascular cartilage tissue. Additionally, the biomechanical signals received by the viscoelastic ECM are transduced and propagated to the resident chondrocytes. The molecular ef-

fects of compressive forces on gene expression in chondrocytes has been stud- ied using several different models, largely ex vivo, including cartilage explants

(3-8), and in vitro systems such as agarose (9-11), alginate (5) and other poly- meric scaffolds (12-14).

Although the cartilage explant tissue models provide exceptional physiological simulation of in vivo biomechanics, its tissue level environment and signaling complexity creates challenges when dissecting the individual molecular responses to the applied compressive forces. The gene expressions in chondrocytes under compression are dependent on the magnitude (3,8,12), frequency

(10) and duration (5,6,8,10,12,13,15) of applied compressive forces. Dynamic compression of cartilage explant cultures at low magnitudes (3–5% strain), whether intermittent or periodic, leads to cyclic changes in pressure, deformation, and fluid flow in the cartilage. At the molecular level, the gene expression of Cartilage Oligomeric Matrix Protein (COMP) in chondrocytes under dynamic compression has revealed that direction (uniaxial or multiaxial), orienta- tion (rotation about cylindrical axis), periodicity (oscillation of scaffold) of the applied forces, as well as culture period, impact the transcriptional profile of chondrocytic gene expression (14). Dynamic compression has been shown to up-regulate expression of anabolic genes such as Aggrecan/ACAN, Collagen

Type II/COL2A1, and Tissue Inhibitor of Metalloproteinase 3 (TIMP3) (15), while down-regulating specific genes of the Matrix Metalloproteinase (MMP) family (4,9,15,16). More importantly, compressive forces at low magnitudes

have been shown to be anti-inflammatory in nature, as evidenced by the findings that when chondrocytes seeded into agarose hydrogel scaffolds are stimulated by exogenous IL-1β, sinusoidal form compression at 15% strain was shown to suppress Aggrecanase 1 (ADAMTS4) and Aggrecanase 2

(ADAMTS5), but not MMP3 gene expression, as well as prevent the downregulation of aggrecan in the presence of inflammation (3,6). Similarly, compressive forces inhibit interlekin (IL)-1β-induced nitric oxide synthase 2A

(iNOS/NOS2A) and cyclooxygenase 2 (COX2/PTGS2) expression, (17, 18) and up-regulate proteoglycan synthesis and cell division in the presence or absence of IL-1β (Figure 2) (19, 20).

Interestingly, exposure of chondrocytes to dynamic compression of higher magnitudes (2 MPa) for as little as 1 h can be traumatic and induces generation of nitric oxide (NO), COX2/PTGS2, prostaglandin E2 (PGE2), MMP1, and inhibits proteoglycans and decorin (DCN) mRNA expression (21-23). Similarly, long term (24–48 h) compressive strain of 25%–50% is catabolic, and substantially down-regulates expression of aggrecan and COL2A1, and up-regulates

MMP3, MMP9, MMP13, and ADAMTS4 expression to induce cartilage destruction (16, 24).

Contrary to dynamic loading, static compression induces an initial accumulation of interstitial hydrostatic pressure within the cartilage, however, this pressure reaches an equilibrium stasis due to stress relaxation of the tissue (3). 6

Static compression invariably down-regulates anabolic gene expression

(4,12,16) and up-regulates catabolic (MMP3, MMP9, MMP13, ADAMTS4) (4) and inflammatory (TNF-α, COX2/PTGS2, iNOS/NOS2A) (4,6) gene expressions. Murata et al. have shown that static compression activates the IL-1 signaling pathway by upregulating IL-1α (IL1A), IL-1β (IL1B) and NOS2A gene expressions in the presence and absence of IL-1 receptor antagonist (Figure 2)

(6).

At cellular level, the presence of a developed pericellular matrix appears to be essential for the appropriate mechanotransduction of biomechanical forces into gene expression in chondrocytes (25). Giannoni et al. showed the up-regulation of COMP gene expression in either alginate or cartilage explant cultures under long-term dynamic compression. However, the mechano-sensitivity required for COMP activation was lost by inhibiting the availability of the transmem- brane receptor protein, β1 integrin (ITBG1), possibly by disabling fibronectin- dependent signaling (26). These findings emphasize the requirement of opera- tive cell-matrix signaling in order to achieve chondrocyte transcriptional activation.

Currently, there is little known about the mechanisms by which compressive forces are converted into intracellular biochemical events. Beta-integrins (Fi- bronectin Receptors, CD29) are the mechano-receptors responsible for trans- ducing many biomechanical signals to the cytoplasm, and the biomimetic pep-

tide-amphiphile, GRGDSP, abolishes the integrin-mediated compression- induced chondrocytic responses to compressive forces (19, 27, 28). Recently, investigators have shown that the biomechanical signaling is processed at the level of nuclear envelope deformation (29).Both static and dynamic compressive forces induce intracellular actin and vimentin reorganization, however, the significance of this reorganization has yet to be fully defined (30).

Regulation of inflammation and Repair in Chondrocytes by Tensile Strain

During joint movement, in addition to compressive forces, chondrocytes in cartilage are exposed to tensile and shear forces. Examination of the tensile and compressive properties of cartilage has shown that fluid-flow dependent viscoelasticity dominates the compressive response of cartilage, whereas intrinsic solid matrix viscoelasticity dominates the tensile response. Furthermore, the dynamic compressive modulus of cartilage is critically dependent upon ele- vated values of the dynamic tensile modulus, suggesting that tensile forces are generated during joint movement (2).

Early investigations of the responses of chondrocytes to tensile forces applied in vitro showed that similar to compressive forces, chondrocytes perceive and respond to tensile forces in a magnitude-dependent manner. At low magnitudes, tensile forces act as potent anti-inflammatory signals and inhibit IL-1β-,

TNF-α-, and LPS-induced pro-inflammatory gene transcription (2, 31, 32).

However, at high / hyperphysiologic magnitudes these signals act as traumatic signals and induce production of pro-inflammatory mediators such as NOS2A,

COX2, MMPs, and NO, irrespective of the presence of an inflammatory stimulus (21). One of the pathways regulated in chondrocytes following mechanical stimulation is the signaling cascade involved in the inflammatory responses.

Dynamic tensile strains of low magnitudes (2.5–7.5% elongation) significantly suppress IL-1β- and TNF-α-dependent INOS/NOS2A, COX2/PTGS2, MMP13 and MMP1 expression, as well as PGE2 and NO production in articular chondrocytes (32-36). In parallel, cyclic tensile strain also inhibits cartilage degradation by up-regulating the mRNA expression of TIMP2 and COL2A1 following their IL-1β-dependent suppression (37). Furthermore, cyclic tensile strain could augment cartilage repair by facilitating the induction of Aggrecan mRNA and attenuation of IL-1β-induced suppression of proteoglycan synthesis (Figure

2).

In subsequent reports, elaboration of the findings on the anti-inflammatory effects of dynamic tensile strain in fibrochondrocytes of the temporomandibular joint (TMJ) and knee meniscus have also shown inhibition of IL-1β- and TNF-

α-induced iNOS/NOS2A, COX2/PTGS2, and MMP1 mRNA expression, NO and PGE2 production, and MMP1 protein synthesis. Extensive in vivo evidence has shown that proteoglycan synthesis is an important regulator of cartilage repair and demonstrated that the application of exogenous proteoglycans can prevent cartilage loss in arthritic joints. In parallel to inhibiting pro-inflammatory 9

gene transcription, cyclic tensile strain also abrogates IL-1β-induced downregulation of proteoglycan, COL2A1, and TIMP2 synthesis (38, 39). Thus, tensile forces of low magnitudes are not only anti-inflammatory, but are also reparative in their actions.

Continuing investigations on the role of cyclic tensile strain on the expression and regulation of MMPs in fibrochondrocytes of the TMJ demonstrated a significant increase in mRNA expression and synthesis of MMP3, MMP7, MMP8,

MMP9, MMP13, MMP16, and MMP17 in response to IL-1β. This IL-1β- induced up-regulation of MMPs was significantly inhibited by the application of dynamic tensile strain. However, MMP2, MMP11, MMP14, TIMP1, TIMP2 and TIMP3 mRNA expressions were not affected by either IL-1β-induced inflammation or the application of tensile strain (40). Interestingly, studies on the expression of pro-inflammatory mediators in fibrochondrocytes from the TMJ or knee meniscus showed that, 5–20% magnitude of tensile strain did not produce any pro-inflammatory effects, while 15% tensile strain was optimal for more than 90% suppression of the above mentioned pro-inflammatory mediators. This suggested that mechanical signals inhibit proinflammatory gene expression in both chondrocytes and fibrochondrocytes, while the threshold of their sensitivity differs with the cell type.

Mechanical signals also regulate expression of Tumor Necrosis Factor (TNF)

Superfamily Member 11 (TNFSF11/RANKL), TNF Receptor Superfamily 10

Member 11A NFKB activator (TNFRSF11A/RANK), and TNF Receptor Super- family Member 11B (Osteoprotegerin) (TNFRSF11B/OPG) in fibrochondrocytes of the rat meniscus. While IL-1β increases the expression of

RANKL/TNFSF11 and RANK/TNFRSF11A, cyclic tensile strain suppresses this expression (39). Using the same biomechanical and inflammatory model,

TNFRSF11B/OPG expression was unaffected by the application of cyclic tensile strain. These effects of cyclic tensile strain were also found to be both magnitude- and frequency-dependent (39).

Dynamic tensile forces inhibit inflammation in a sustained manner

The findings that the application of dynamic tensile strain can cause a rapid inhibition of the actions of IL-1β and TNF-α in chondrocytes in vitro led to the inquiry of how long the actions of mechanical signals last in suppressing proinflammatory gene inductions. In this model of biomechanical strain, chondrocytes were subjected to cyclic tensile forces for the initial 1.5 h, 4 h, or 8 h of the experimental design, while being exposed continuously to an IL-1β inflammatory environment over a 24 h time course. These results showed that the suppressive effects of biomechanical strain were sustained over various time intervals of application during a 24 h experimental time course, and as short as

90 min of tensile forces were sufficient to inhibit IL-1β-induced COX2/PTGS2, iNOS/NOS2A, MMP9, and MMP13 expression for the ensuing 8 h, despite the continued presence of a pro-inflammatory environment [32, 40]. Similarly, an 11

8 h exposure of dynamic tensile strain of low magnitudes was sufficient to inhibit pro-inflammatory gene induction for the ensuing 16 h, but was not sufficient to inhibit iNOS/NOS2A expression for an additional 28–40 h. These observations suggested that the anti-inflammatory cascade initiated by the application of dynamic tensile forces persisted despite the cessation of the biomechanical stimuli.

Dynamic tensile strain regulates the NF-κB pathway to induce or inhibit pro-inflammatory gene transcription

The preceding findings that dynamic tensile strain can cause rapid induction or inhibition of pro-inflammatory genes in a magnitude- and frequency-dependent manner provided the essential foundation for further exploration of pathways through which the actions of mechanical signals are converted into functional responses. In order to extend these observations, attention soon turned to the

NF-κB signaling pathway as a possible link between tensile loading and chondrocytic responses to pro-inflammatory cytokines. First, the NF-κB signaling cascade is the major pathway that controls pro-inflammatory gene transcription. Second, both low and high magnitudes of tensile strain regulate proinflammatory gene transcription. Finally, the effects of tensile forces are not mediated by the immediate down-regulation of IL-1 or TNF-α receptors on chondrocytes. Together, these findings suggested that chondrocytic responses to mechanical loading occur largely independent of traditional cytokine recep- 12

tors, and that biomechanical activation of the NF-κB signaling pathway is essential for the propagation of these actions.

In order to identify the key target molecule(s) that are regulated by dynamic tensile forces, it is essential to understand the many roles of NF-κB and its complex web of intracellular regulation. This knowledge provides the funda- mental rationale for focusing on IL-1β-induced NF-κB nuclear translocation and the prerequisite upstream events involved in its activation in response to dynamic biomechanical stimuli. In chondrocytes, biomechanical signals are perceived in a magnitude- and frequency-dependent manner to promote or attenuate pro-inflammatory gene transcription [37]. Biomechanical signals are transduced to cells by surface molecules such as β-integrins and focal adhesion kinases (FAKs/PTK2 kinases) [41] The pro-inflammatory response exhibited by articular chondrocytes subjected to tensile strain of higher magnitudes is paralleled by an increased NF-κB nuclear import, providing support for a central role of NF-κB in the pro-inflammatory signals generated by tensile loading

[38]. Conversely, at lower magnitudes, biomechanical signals inhibit nuclear translocation of NF-κB transcription factors, and act as potent inhibitors of IL-

1β- and TNF-α-dependent pro-inflammatory gene transcription [37,42,43].

Multiple cytokine-induced proinflammatory pathways converge at the signalosome comprised of Inhibitor of Kappa Light Polypeptide Gene Enhancer in B

cells Kinase, IKKα (IKKA/IKK1/CHUK), IKKβ (IKKB/IKK2/IKBKB) and

IKKγ (NEMO/IKBKG) to activate downstream events in the NF-κB cascade.

Upon phosphorylation by IKKs, Inhibitor of Kappa Light Polypeptide Gene

Enhancer in B cells proteins, I-κB (NFKBI), are ubiquinated and marked for proteosomal degradation. The liberation of NF-κB from I-κB complexes is followed by its phosphorylation at multiple sites in a stimulant-dependent manner and eventual translocation into the nucleus. The binding of NF-κB to its consensus sequences leads to transcription of a variety of genes including proinflammatory cytokines and mediators, as well as several of the molecules required for the activation of NF-κB signaling cascade. This classical model of

NF-κB activation by TNF-α or IL-1β is well documented, and its complexity evolves from its regulation at multiple intracellular levels, in a cell-dependent as well as stimulus-dependent manner

Characterization of anti-inflammatory actions of dynamic tensile strain has shown that it intercepts multiple steps along the NF-κB signaling cascade to block IL-1β-induced pro-inflammatory gene transcription. As an initial step in regulating NF-κB signaling, dynamic tensile strain markedly abrogates IL-1β- dependent IKK activation, leading to drastic reductions in the phosphorylation and subsequent degradation of I-κBα (NFKBIA) and I-κBβ (NFKBIB). Conse- quently, NF-κB remains sequestered in the cytoplasm by I-κBα and I-κBβ.

This results in the suppression of NF-κB nuclear translocation and repressed

transcriptional activation of pro-inflammatory genes. Additionally, signals generated by tensile strain abrogate the IL-1β-induced increase in I-κBα and I-

κBβ gene expression to unstimulated control levels. While I-κBα expression is controlled by NF-κB transcriptional activity, I-κBβ has not been shown to be under the transcriptional control of NF-κB [44]. Since the pathways that culmi- nate in I-κBβ resynthesis after its degradation are incompletely characterized, the main conclusion of these findings is that the signals generated by cyclic tensile strain likely interact with proteins other than those controlled by NF-κB to regulate pro-inflammatory gene induction. Finally, dynamic tensile strain induces a rapid up-regulation of I-κBα shuttling from the cytoplasm into the nucleus to faciliate the export any translocating NF-κB so as to prevent/terminate its transcriptional activity. Thus, the collective actions of cyclic tensile strain at multiple regulatory levels within the NF-κB signal transduction pathway prevent the activation of NF-κB transcription factors and ultimately result in the inhibition of pro-inflammatory gene transcription [45]

Shear forces, chondrocyte metabolism, and gene expression

Studies have shown that fluid-induced shear stress regulates chondrocyte metabolism. Articular chondrocytes exhibit a dose- and time-dependent response to shear stress that results in the release of soluble mediators and extracellular matrix macromolecules [46]. Additionally, Lee et al. have shown that exposure

to shear stress induces NO release in both a dose- and time-dependent manner.

Shear stress-induced NOS2A mRNA expression and NO production have been shown to be associated with decreased mRNA expressions for the cartilage matrix proteins ACAN, and COL2A1 [47]. Similarly, fluid shear forces have been shown to induce COX2/PTGS2 mRNA expression and suppress PI3K activity, which in turn is associated with modulation of apoptotic and antioxidant pathways [48]. On the contrary, Fitzgerald et al. have shown an up-regulation of the transcription of proteoglycans and COL2A1 by cyclic shear stress [4]. These findings suggest that the biomechanical signals introduced by the fluid shear deformation of chondrocytes and cartilage are an additional important component of cartilage homeostasis.

CHAPTER 2

DYNAMIC BIOPHYSICAL STRAIN MODULATES PRO- INFLAMMATORY GENE INDUCTION IN MENISCAL FIBROCHONDROCYTES

The meniscus of the knee, positioned between the articular surfaces of the femur and tibia is crucial for the biomechanical stability of the joint. The geometry, type of extracellular matrix, and cells that synthesize the matrix of this fibrocartilage are well adapted to withstand repetitive biomechanical forces under physiological conditions (51, 71). The fibrochondrocytes of this cartilage synthesize matrix that consists of type I and type II collagens in addition to glycosaminoglycans to provide increased ability to endure tensile, compressive, and torsional forces during joint movement (50, 58, 66). Biomechanical forces regulate fibrochondrocyte functions in a complex manner (49, 51, 64, 66), and any changes in the magnitudes of these forces elicit profound effects (37, 57, 16, 64, 68, 73).

Non-physiological or traumatic loading of the joints can be injurious and induce upregulation of pro-inflammatory mediators like nitric oxide (NO), inducible NO synthase (iN-

OS), matrix metalloproteinase-1 (MMP-1), as well as down regulation of matrix proteins like proteoglycans, decorin, and collagen type II, in vitro (56, 57, 16, 69, 22). Similarly,

in vivo, increased stress or damage to meniscus can lead to progressive degenerative

changes leading to osteoarthritis and associated joint dysfunction, suggesting a role

of mechanical signals in the pathogenesis of arthritic diseases (60, 62, 16, 67, 68).

The responses of fibrochondrocytes to biomechanical forces are little understood.

However, dynamic mechanical stimulation of appropriate / physiological magnitudes is

shown to significantly influence anabolic activities in articular chondrocytes (51). The

support that these signals are reparative in nature comes from studies demonstrating that

mechanical signals of low/physiological magnitudes are potent anti-inflammatory signals

and inhibit IL-1β and TNF- α-induced proinflammatory gene induction, induce matrix

synthesis, and prevent dedifferentiation of chondrocytes (49, 54, 35, 63, 64, 73, 36). The

fibrocartilage of the meniscus is notorious for its limited regenerative capacity, and there-

fore, salvage of damaged meniscus has been a major focus of clinical as well as experi-

mental studies. Biomechanical signals in the form of active or passive motion to improve

healing of the fibrocartilage have been tried with variable success (52, 53, 59, 61). Re-

cently, in a model of experimental antigen-induced arthritis, the integrity of meniscal car-

tilage is shown to be better preserved in joints exposed to controlled biomechanical load-

ing (37, 59, 65). These observations suggest that appropriate mechanical forces can be

used for the repair and rehabilitation of meniscal cartilage.

To develop appropriate therapeutic strategies for the repair or regeneration of damaged

fibrocartilage, in this study we have made an attempt to understand the mechanisms of

actions of biomechanical signals on the cells of fibrocartilage. It was our hypothesis that

fibrochondrocytes of the meniscus respond to mechanical signals in a magnitude- and

frequency-dependent manner. Furthermore, biomechanical signals suppress inflammation

by inhibiting proinflammatory gene transcription, and these effects are sustained for ex- tended periods of times. We show that signals generated by dynamic tensile forces play

an important role in transcriptional regulation of proinflammatory genes in fibrochondro-

cytes.

Methods

Cell harvesting and culture

Menisci harvested from the knees of female Sprague Dawley rats (200-250 g; Harlan, IN)

were minced and treated with 0.2% trypsin for 10 min, followed by 0.15% collagenase I

(Worthington NJ) digestion in a two-compartment digestion chamber for 2 h. Fibrochon-

drocytes released were washed in TCM [50% DMEM/50% F-12 medium (Cellgro®, Me- diatech, VA) supplemented with 10% defined Fetal Calf Serum (Hyclone, Logan, UT) and penicillin/streptomycin (100 U/100 μg/ml)], and cultured at 37ºC and 5% CO2. Sub- sequently, the phenotype of meniscal cells was confirmed by the synthesis of collagen

type-I and -II, aggrecan, biglycan and versican. The phenotype of meniscal cells was

found to be stable during the first 3 passages.

Exposure of cells to dynamic tensile forces

Fibrochondrocytes (105/well) were transferred to collagen I coated six well Bioflex-II

plates (Flexercell Int, NC), and grown for 5 days to achieve 70 to 80% confluence. The

medium was replaced with TCM containing 1% FCS, 20 h prior to subjecting cells to

DTF at various magnitudes (3, 6, 9, 12, 15, or 20%) and frequencies (4, 25, 50, 100 or

250 mHz) in a Flexercell strain unit (Flexcell Int, NC) that exposes cells to equibiaxial

tensile forces. In all experiments the load was given in a square waveform and remained

the same at all frequencies and magnitudes (Fig. 1A). Cells were subjected to four different treatment regimens: 1) Untreated cells, 2) cells exposed to recombinant human inter-

leukin-1β (IL-1β; 1 ng/ml, Calbiochem, CA), 3) cells exposed to DTF alone, and 4) cells

exposed to both DTF and rhIL-1β (1 ng/ml). IL-1β at 1 ng/ml reproducibly induced iN-

OS mRNA expression and NO production (49, 35, 36).

RNA purification and real-time polymerase chain reaction

RNA was isolated from cells with RNeasy kit (Qiagen, CA) according to manufacturer’s

recommended protocols. Briefly, cells were scrapped in a total of 350 μl of RLT buffer,

and DNA shredded by passing through a QIAshredder (Qiagen, CA) column and spin-

ning at 13,000 x g for 2 min. Subsequently, the cell extracts were mixed with an equal 20

volume of 70% ethyl alcohol, treated with DNAse, and loaded on a silica gel column to bind RNA. The RNA bound to silica gel was thoroughly washed, and eluted with 50 μl of sterile water. The concentration of RNA was assessed by reading absorbance at 260/280 nm in a Biophotometer 6131 (Eppendorf, Germany).

Gene specific primer sequences were selected using the Taqman Probe and Primer

Design function of the Primer Express v1.5 software (Applied Biosystems, CA). The sense and antisense sequences of rat primers used are as follows: MMP-13 sense 5’-

GTTCAAGGAATCCAGTCTCTCTATGG-3’, antisense 5’-

TGGGTCACACTTCTCTGGTGTTT-3’, probe 6-

FAMd(CCAAGGAGATGAAGACCCCAACCCTAAGC)BHQ-1 (XM343345); TNF-_ sense 5’-CCCAGACCCTCACACTCAGATC-3’, antisense 5’-

TCGTAGCAAACCACCAAGCAG-3’, probe 6-

FAMd(TCGAGTGACAAGCCCGTAGCCCA)BHQ-1(X66539); iNOS sense 5’-

TTCTGTGCTAATGCGGAAGGT-3’, antisense 5’-GCTTCCGACTTTCCTGTCTcA -

3’, probe 6-FAMd(CCGCGTCAGAGCCACAGTCCT)BHQ-1 (D44591). Reverse transcription reactions were carried out using 2 μg RNA and TaqMan Reverse Transcription

Reagents, followed by real-time PCR using Taqman PCR Master Mix and ABI Prism

7700 Sequence Detection System (Applied Biosystems, CA). Reactions were performed as follows: Cycle I (1X): 95º C for 3.0 min, Cycle II (50X): Step 1 at 95º C for 0.3 min, followed by Step 2 at 55º C for 0.3 min, and Step 3 at 72º C for 30 min, Cycle III at 40º C hold. Following amplification, melt curve was obtained to ensure that primer-dimers or non-specific products had been eliminated or minimized. The data obtained by real-time

PCR, was analyzed by the comparative threshold cycle (CT) method. In this method, the

amount of the target, normalized to GAPDH, and relative to a calibrator (either untreated sample or IL-1β-stimulated cells), is given by 2-∆ ∆ CT, where ∆CT = ∆CT (sample) - ∆CT

(calibrator), and ∆CT is the CT of the target gene subtracted from the CT of GAPDH.

Semiquantitative analysis of proteins by Western blot analysis

The semiquantitative estimation of proteins was carried out by Western blot analysis (49,

35).Briefly, cells exposed to various regimens of mechanostimulation were lysed in ice

cold Tris buffered saline (TBS) containing protease inhibitor cocktail (Roche, IN) and the

extracted proteins were loaded on the SDS-10% acrylamide gels. The proteins were elec-

trophoretically transferred to nitrocellulose membranes (Bio-Rad, CA) and identified by

monoclonal mouse anti-MMP-13 IgG (1:1000 dilution; Calbiochem, CA), mouse anti-

iNOS IgG (1:200 dilution; BD Bioscience, CA), rabbit anti-TNF IgG (1:500 dilution;

Biosource, CA). Monoclonal mouse anti-β-actin IgG (1:20,000; Abcam, MA) was used

to reprobe the same blots as a standard. Horseradish peroxidase labeled donkey anti-

mouse IgG (1:10,000 dilution; Chemicon, CA) or goat anti-rabbit antibody (1:10,000 di-

lution; Santa Cruz) was used as secondary antibody. The presence of HRP was detected

by Luminol (Amersham, IL). The semiquantitative expression of proteins was estimated

in digital images of luminescence in bands of Western blots by Kodak Image Station

1000 and Kodak 1D image analysis software (Kodak, CT). Expression of β-actin

was used to standardize the protein input in each lane of a blot (49, 36). The relative luminescence of bands in cells treated with IL-1β and DTF was estimated as percent of luminescence in the bands from cells treated with IL-1βalone.

Measurements of Nitric Oxide

NO in the culture supernatants of fibrochondrocytes was quantified by modified Griess reaction as described earlier (49, 35).

Results

Mechanosensitivity of meniscal fibrochondrocytes to DTF at various magnitudes

Prior to determining the role of DTF on fibrochondrocytes, we examined the effects of

0.1 ng/ml to 25 ng/ml concentrations of IL-1β on the induction of NO in fibrochondrocytes. The optimal NO production was observed at 1 ng/ml, beyond which a plateau in

NO production was observed (data not shown). In order to examine the effects of DTF on fibrochondrocytes, these cells were exposed to equibiaxial strain at various magnitudes and 0.05 Hz in the presence or absence of IL-1β. Fibrochondrocytes did not express iN-

OS mRNA constitutively. Exposure of fibrochondrocytes to IL-1β (1 ng/ml) exhibited a marked upregulation of iNOS mRNA and protein induction (Fig. 1 A, B). Meniscal cells 23

when exposed to DTF alone at 5%, 10%, 15%, or 20% did not exhibit iNOS mRNA expression (data not shown). However, in the presence of IL-1β, DTF inhibited IL-1β- induced iNOS expression in a magnitude-dependent manner: as little as 5% equibiaxial strain inhibited 57%, whereas 15% DTF suppressed more than 98% of IL-1β-induced iNOS mRNA expression (Fig. 1 A).

In parallel experiments, Western blot analysis revealed that DTF also significantly inhibited iNOS synthesis in a dose dependent manner, exhibiting an inhibition of 42% iNOS synthesis by 5% DTF and more than 95% iNOS synthesis by 15% DTF (Fig. 1 B). The confirmatory studies analyzing NO production in culture supernatants of cells exposed to

DTF for 36 h also showed a significant inhibition of IL-1β-induced NO production parallel to inhibition of iNOS synthesis (Fig. 1 C).

Anti-inflammatory effects of DTF are frequency dependent

In the next series of experiments, we investigated whether the inhibitory effect of DTF on the IL-1β-induced iNOS mRNA expression is frequency dependent. Because DTF at a magnitude of 15% exhibited suboptimal suppression of IL-1β-dependent iNOS synthesis and NO production, the experiments were conducted at a magnitude of 15%. As shown in

Fig 1 D, the frequency of DTF is as important in regulating proinflammatory gene expression, as is magnitude. A frequency as rapid as 0.25 Hz significantly inhibited more

than 48± 6% IL-1β-induced iNOS mRNA expression. Similar to the magnitudes of DTF,

the effects of frequencies of biomechnical signals were also dose-dependent. A frequency

of 0.025 Hz inhibited IL-1β-induced iNOS expression maximally. However, exposure of

cells to decreasing frequencies of strain versus rest demonstrated that 0.004 Hz minimally

inhibited IL-1β-induced iNOS gene expression (Fig. 1 D).

Dynamic tensile strain blocks IL-1_-induced synthesis of proinflammatory media-

tors

Because IL-1β-upregulates multiple proinflammatory genes (60, 62), we next determined

whether the effects of DTF at a magnitude of 20% DTF and a frequency of 0.05 Hz also

blocked the synthesis of other proinflammatory proteins. Tumor necrosis factor (TNF)-α

and MMP-13 representing 2 major categories of proteins, a cytokine and a matrix metal-

loproteinase involved in cartilage destruction (63, 64) were selected. Control cells and

cells exposed to DTF alone for 4 or 24 h did not express mRNA for TNF-α or MMP-13.

However, meniscal cells exposed to IL-1β exhibited a significant upregulation of TNF-α and MMP-13 induction. Quantitative analysis by real time PCR revealed that co-exposure of cells to DTF for 4 or 24 h abolished more than 95% IL-1β-induced TNF-α and MMP-

13 (Fig. 2 A). As expected, in parallel experiments the synthesis of TNF-α and MMP-13 was also suppressed by the actions of DTF in IL-1β treated cells after 24 h (Fig. 2 B and

C).

The suppression of proinflammatory gene induction by DTF is sustained despite the

presence of IL-1β

The observations that DTF suppresses IL-1β-induced proinflammatory gene transcription in fibrochondrocytes, prompted us to investigate the temporal regulation of IL-1β-

induced genes by biomechanical signals. In these experiments, IL-1β was added to fibro-

chondrocytes at the start of the experiment and remained in the medium for 24 h. The

cells were exposed to DTF (15%, 0.05 Hz) for the first 4, 8, 12, 16, or 20 h followed by a

period of rest for 20, 16, 12, 8, or 4, respectively. All cultures were analyzed at 24 h for

the expression of iNOS mRNA. Additionally, in presence of IL-1β cells were exposed to

DTF continuously for 4, 12, 16, or 20 h, and analyzed immediately after DTF exposure.

As evident in Fig. 3 A, in the presence of IL-1β, continuous exposure of cells to DTF for

4, 12, 16, or 20 h exhibited significant suppression of IL-1β- induced iNOS gene induc-

tion. On the contrary, analysis of cells exposed to DTF for 4 h followed by 20 h of rest

revealed that DTF suppressed more than 60% of IL-1β-induced iNOS expression.

Analogous to these observations, DTF exposure for 8 h strongly suppressed (96% ± 5%)

IL-1β- induced iNOS gene induction during the ensuing 16 h (Fig 3 B). Surprisingly, ex-

posure of cells to DTF for 12, 16, or 20 h, followed by rest, failed to block the IL-1β- induced iNOS gene expression, when analyzed following 12, 8, or 4 h of rest. As expected, continuous exposure of fibrochondrocytes to DTF for 24 h blocked the iNOS mRNA expression induced by IL-1β. Thus, the anti-inflammatory effects of DTF are sustained following short-term exposures of fibrochondrocytes to DTF, whereas long-term exposures of DTF followed by rest did not block the actions of IL-1β in a sustained manner. On the other hand, continuous exposure of cells to DTF consistently blocks the actions of IL-1β (Fig 3).

Discussion

The findings of this study support the hypothesis that biomechanical signals regulate the proinflammatory gene expression in fibrochondrocytes of the meniscus in a magnitude and frequency dependent manner. In this in vitro system, we have examined the responses of meniscal fibrochondrocytes to two factors, IL-1β that is involved in the pathogenesis of the meniscus and dynamic strain that is known to elicit rehabilitative effects on the cartilage (49, 51, 52, 35, 55, 37, 59). The effects of IL-1 are mediated by NO, therefore the expression of iNOS and resultant NO production was used as a marker of inflammation (57, 70). Signals generated by DTF of 5% to 20% are not perceived as traumatic signals in meniscal fibrochondrocytes. Contrarily, DTF at magnitudes of 5 to 27

20% are anti-inflammatory and suppress IL-1β-induced iNOS mRNA expression and

consequently its synthesis and NO production in a magnitude dependent manner. In these

experiments, DTF blocked mRNA expression that was paralleled by inhibition of protein

synthesis, suggesting that mechanical signals inhibit IL-1β-induced proinflammatory

gene induction at transcriptional level. Present technical limitations imposed by

the strain devices restrict the examination of the effects of magnitudes of DTF higher

than 20%.

These findings are similar to those observed in articular chondrocytes where tensile as

well as compressive forces of low magnitudes are shown to exert anti-inflammatory ef-

fects (54, 57).However, fibrochondrocytes exhibit suppression of proinflammatory gene

induction at 2 to 3 fold higher magnitudes of mechanical forces than articular chondro-

cytes (49, 35, 36).Fibrochondrocytes serve discrete functions in a diarthroidal joint. In

this cartilage,fibrochondrocytes constitute more than 20% of the tissue, and synthesize

tough intercellular matrix comprising 20-25% collagens type I and II, and only 0.6% to

0.8% glycosaminoglycans (GAGs) (50, 51, 58, 66, 27). While this collagen-rich stiffer matrix provides the ability to undergo tensile, compressive, torsion, and shear forces during loading, its more than ten-fold lower GAG content may provide limited ability for hydrodynamic shock absorption.

This may therefore lead to exposure of meniscal chondrocytes to higher magnitudes of

biomechanical forces during normal joint function, and their ability to withstand 2 to 3

fold higher magnitudes of mechanical forces without perceiving them as a proinflamma-

tory signal (72). The frequency of mechanical signals is as important in regulating the responsiveness of cells as its magnitude. Mechanical forces of rapid frequencies (0.25 and 0.025 Hz) effectively inhibit IL-1β-induced proinflammatory responses, whereas a

slower frequency (0.004 Hz) is less effective in revoking the IL-1β actions in meniscal

fibrochondrocytes. While earlier findings have shown that static forces are proinflamma-

tory to cartilage (68, 22), the present findings suggest that signals at lower frequencies

inhibit IL-1β-induced proinflammatory gene transcription but to a lesser extent than

higher frequencies, in fibrochondrocytes.

Injury or trauma to cartilage induce a 250 fold increase in the production of MMPs, and

several fold increase in the expression of cytokines and NO production (16, 68). Me-

chanical signals, despite the presence of IL-1β in the medium, block the expression of

proinflammatory mediators involved in cartilage destruction. For example, TNF-α, the major mediator involved in the pathogenesis of rheumatoid and osteoarthritis, is upregulated in response to an injury (16, 68). DTF effectively inhibits synthesis of TNF-α by

suppressing expression of its mRNA in the presence of exogenous IL-1β. Similarly,

MMP-13, by initiating cleavage of collagens, is one of the major matrix-associated pro- teinase involved in the development of arthritic lesions. Downregulation of MMP-13 syn-

thesis by DTF represents another mode of protective effects of biomechanical signals on inflamed joints. In this respect, the actions of DTF are similar to the inhibitors of inflammation known to be palliative to inflamed joints (51, 55), and suggest that DTF likely serves as a key signal in preserving the functional integrity of the inflamed fibrocartilage. Proinflammatory gene transcription by IL-1β is mediated by nuclear factor-κB

(NF-κB) transcription factors, which activate a plethora of proinflammatory gene expression (49).

Additionally, the down-regulation of proinflammatory responses by biomechanical signals has been shown to be mediated by inhibition of nuclear translocation of NF-_B in articular chondrocytes (49). Therefore, it is likely that inhibition of IL-1β-induced proinflammatory protein induction by biomechanical signals in fibrochondrocytes is also brought about by inhibition of nuclear translocation of NF-κB. In such a case, it is tempt- ing to speculate that biomechanical signals may inhibit number of other proinflammatory genes controlled by NF-κB. The role of this pathway in the anti-inflammatory actions of biomechanical signals needs further elucidation.

The upregulation of proinflammatory molecules following a traumatic injury is sustained up to 72 h following its initiation (16). The experiments to examine the time-dependent effects of DTF have revealed that in the constant presence of IL-1β, (i) exposure of DTF for 4 or 8 h is sufficient to down-regulate iNOS induction for ensuing 20 or 16 h, (ii) exposure of cells to DTF for 12, 16, or 20 h minimally suppresses IL-1β-induced iNOS in-

duction when examined after a rest of 12, 8, or 4 h respectively, and (iii) the effects of

DTF are sustained from 4 to 24 h if DTF is applied to cells continuously. These observations suggest that long lasting anti-inflammatory effects of DTF can be achieved in cells by exposure of cells from 4 to 8 h of DTF. Similarly, constant exposure of cells to DTF without rest results in a sustained suppression of proinflammatory responses. However, exposure of longer durations of DTF (12, 16 or 20 h) followed by a rest (8 or 4 h) does not have the similar sustained effects on iNOS gene suppression. Whether this failure of longer durations of DTF followed by rest to repress iNOS induction is due to cellular fatigue following long exposures to mechanical signals is yet to be elucidated.

On this note, motion-based therapies in the form of continuous passive motion are known to be beneficial to inflamed joints and reduce pain and inflammation (52, 53, 59,

61). Our data also suggest that constant exposure to mechanical signals suppresses the IL-

1-induced proinflammatory gene induction in a sustained manner. Furthermore, limited exposures to DTF can elicit long lasting effects on the inhibition of proinflammatory gene induction. However, there is an optimal period of time that is sufficient for the effectiveness of biomechanical signals, beyond which the biomechanical signals become less effective. Whether therapies similar to CPM can also be utilized for shorter intervals without losing their effectiveness is yet to be determined.

Overall, this is the first evidence that dynamic tensile strain by blocking proinflammatory

gene transcription acts as an anti-inflammatory signal on fibrochondrocytes of the meniscus. The mechanoresponsiveness of fibrochondrocytes varies according to the magnitude and frequency of the signals generated by the tensile forces. More importantly, the effectiveness of anti-inflammatory actions of biomechanical signals is sustained despite the surrounding proinflammatory environment but is dependent on the length of their exposure to cells. These in vitro findings, may provide the molecular basis for the beneficial effects of motion-based therapies on inflamed joints. Further examination of these effects in vivo is essential for translating anti-inflammatory effects of biomechanical signals into successful therapeutic modalities.

Acknowledgements: This study was supported by National Institute of Health grants,

AT000646, DE15399, and HD40939.

Figure 1. Regulation of IL-1_-dependent iNOS induction by dynamic tensile forces at various magnitudes and frequencies in meniscal fibrochondrocytes. (A) Real time PCR analysis showing expression of iNOS mRNA after 4 h exposure to DTF of various magnitudes in the presence or absence of IL-1β. (B) Western blot analysis showing iNOS

synthesis after 24 h exposure to DTF of various magnitudes in the presence or absence of

IL-1β. β-actin was used as an internal control to assess protein input in all lanes. (C)

Analysis of NO in culture supernatants of parallel experiments shown in (B), incubated for 36 h. (D) Real time PCR analysis of mRNA in cells exposed to various frequencies of

DTF at a magnitude of 15% in the presence or absence of IL- 1β, showing quantitative differences in iNOS mRNA expression. The data represent one out of three separate experiments with similar results. In all experiments, expression of mRNA for GAPDH was used as an internal standard. * indicates p<0.05 when compared to cells treated with IL-1β alone.

Figure 2. Effect of DTF on IL-1β-dependent MMP-13 and TNF-α induction in fibrochondrocytes. (A) Real-time PCR analysis of the extent of IL-1β-induced MMP-13 and TNF-α mRNA suppression by DTF in the presence of IL-1β. Fibrochondrocytes were exposed to DTF at a magnitude of 20% and a frequency of 0.05 Hz. Cells were either untreated or treated with IL-1β (1 ng/ml) treatment. GAPDH was used as an internal control to equalize mRNA input. Regulation of IL-1β-induced TNF-α (B) and MMP-13

(C) synthesis by DTF in fibrochondrocytes. Beta-actin was used as an internal standard to assure equal input of proteins in each lane. Proteins in each lane of Western blots were assessed by semi-quantitative densitometric analysis. The data are representative of one out of 3 separate experiments performed in triplicates with similar results. The bars represent mean and standard error of the mean. * represents p< 0.05 as compared to IL-1_ treated cells.

Figure 3. Temporal regulation of IL-1β-induced iNOS induction by dynamic tensile

strain. Fibrochondrocytes were exposed to DTF at a magnitude of 20% and a frequency

of 0.05 Hz for 4, 12, 16, or 20 hrs in the presence of IL-1β. Subsequently, RNA was ex-

tracted and the expression of iNOS mRNA was assessed by real-time PCR (A). (B) To

examine how long the effects of DTF are sustained in cells, fibrochondrocytes were ex-

posed to DTF for 4, 8, 12, 16, 20, or 24 h, and allowed to rest for 20, 16, 12, 8, 4, or 0 h,

respectively, in the continuous presence of IL-1β. RNA expression was assessed by real-

time PCR. The data are representative of one out of 3 separate experiments performed in triplicates. The bars represent mean and standard error of the mean. * represents p< 0.05 as compared to IL-1β treated cells.

CHAPTER 3

BIOMECHANICAL SIGNALS EXERT SUSTAINED

ATTENUATION OF PROINFLAMMATORY GENE

INDUCTION IN ARTICULAR CHONDROCYTES

Introduction

Pathophysiologies associated with joint inflammation impose significant impediments to

normal joint function (74, 75). Because of their notable benefits, patients with arthritic

diseases are invariably prescribed rehabilitative therapies for reducing inflammation and

to improve joint function (24, 76-79). Duration of treatment for such therapies varies be-

tween continuous passive motion for several days to a few hours per day of assisted or

voluntary joint movements (80).

However, no consensus has been achieved whether constant motion over long periods of

time, or precise durations of such therapeutic interventions are required to achieve their

optimal benefits. Despite our increasing understanding of the molecular mechanisms of

cartilage development and repair, little is known about how mechanical signals regulate

cartilage repair. For the optimal use of physical therapies to accelerate cartilage repair it

is desirable to understand the mechanisms of actions of mechanical signals in cartilage repair.

Biomechanical forces are critical for the complex process of cartilage development, homeostasis, and functionality (43, 81-82). As mechanosensitive cells, articular chondrocytes perceive and respond to mechanical signals throughout life. Chondrocytes synthesize a unique extra-cellular matrix and their integrity, to a large extent, depends upon intracellular signals generated in response to biomechanical forces (16, 36, 37, 75, 76, 83).

Chondrocytes maintain a functional balance between cartilage degradation and repair by production of various enzymes, cytokines, and matrix associated proteins. A loss of this functional balance is associated with changes in their phenotypic characteristics which invariably leads to joint degenerative disorders like osteoarthritis and rheumatoid arthritis

(84, 85).

Phenotypically, articular chondrocytes are characterized by their ability to synthesize a specific matrix consisting of type II collagen and glycosaminoglycans (GAGs) that allows them to withstand changes in their mechanical environment. However, this phenotype is pliable and in response to a pathological insult, is modulated to an inflammatory

phenotype (74, 86-88). For example, exposure of articular chondrocytes to an inflamma-

tory stimulus leads to their activation and production of interleukin-1β, tumor necrosis

factor-α (TNF-α), prostaglandin E2 (PGE2), nitric oxide (NO), and matrix metallopro-

teinases (MMPs) (38, 75, 85).

Simultaneously, these cells lose their chondrocytic phenotype and fail to synthesize glycosaminoglycans (GAGs) and collagen type II. This inflammatory phenotype is

responsible for amplification of the immune response, cartilage destruction, and patholo-

gies associated with arthritic diseases (74, 86-88). The phenotypic characteristics of chondrocytes are also regulated by biomechanical forces. Both the magnitude and frequency of biomechanical signals are critical determinants in modulating chondrocytic gene expression (21, 33, 38, 89).

Biomechanical forces of high (traumatic) magnitudes are proinflammatory and induce enhanced expression of inflammatory mediators. Proinflammatory alterations in gene expression lead to enzymatic breakdown of cartilage and inhibition of matrix synthesis (15

16, 21). Trauma caused by mechanical forces of high magnitudes induces rapid synthesis of proteases involved in GAGs and collagen degradation12, 25. Biomechanical signals of

lower magnitudes act as potent anti-inflammatory signals that modulate the inflammatory

phenotype to a chondrocytic phenotype by down-regulating proinflammatory gene ex-

pression and up-regulating matrix synthesis (18, 20, 33, 38, 40, 83, 89, 91). For example,

activation of articular chondrocytes by IL-1β or TNF-α results in a marked upregulation

of proinflammatory gene induction (74, 75, 85). Signals generated by mechanical forces 41

of low magnitudes inhibit IL-1β-induced mRNA transcription and synthesis of potent catabolic genes like inducible nitric oxide synthase (iNOS), cyclooxygenase (COX-2), and

MMPs. Simultaneously, these signals upregulate matrix synthesis by counteracting the proinflammatory inhibition of tissue inhibitors of matrix metalloproteinases (TIMP),

GAG synthesis, and collagen II expression (18, 20, 40, 33, 36, 37, 38). In the same con- text, dynamic motion ameliorates the effects of arthritis and alleviates the pain, by exert- ing potent anti-inflammatory responses generating reparative signals (37, 79, 80).

Mounting evidence indicates that biomechanical forces regulate two main categories of genes, those involved in inflammation and those involved in matrix synthesis (15, 18,

21, 38, 89, 91). Pharmacological therapies like non-steroidal anti-inflammatory drugs

suppress the inflammatory response, but are not known to alter the inflammatory pheno-

type or induce cartilage repair or regeneration. Furthermore, these therapeutics provide

transient relief mostly by inhibiting the cyclooxygenase pathway. In comparison, biome-

chanical signals appear to be more potent in that they inhibit the induction of a plethora

of proinflammatory genes, as well as up-regulate synthesis of matrix associated mole-

cules (18, 20, 33, 36, 38, 91). A major consideration for the efficacious therapeutic use of

biomechanical signals is to know how long the anti-inflammatory and reparative effects are sustained following the removal of the biomechanical stimulus. In this report we have examined the time-dependent consequences of the removal of biomechanical signals on the suppression of proinflammatory gene induction. Our studies show that effects of biomechanical signals are sustained for prolonged periods of time even in a proinflammatory environment.

Materials and Methods

ISOLATION OF ARTICULAR CHONDROCYTES

Articular chondrocytes were obtained from the cartilage of shoulder and knee joints from healthy Sprague Dawley rats (36, 38). Approximately 200 micron of superficial cartilage was shaved from the heads of the long bones and finely chopped. The chondrocytes were enzymatically released by digestion with 0.2% trypsin for 15 min, followed by a 3 h digestion with 0.15% collagenase I (Worthington Biochemical Corp, NJ) in a twocom- partment digestion chamber kept at 37° C. Chondrocytes were cultured in TCM [50%

Dulbecco’s modified Eagle’s medium and 50% Ham’s F-12 medium (Mediatech, VA) supplemented with 10% fetal calf serum (FCS) and penicillin/streptomycin (100 U/100

μg /ml)] at 37°C, in an atmosphere of 5% CO2 for 7 days. Subsequently, chondrocytes

(105 /well) were transferred to collagen-I coated six well Bioflex plates (Flexercell Inter-

national, NC), and grown for 5 days to attain 75% to 80% confluence. Chondrocytes in

passages two and three were used for experimentation, where they exhibited typical phe-

notypic markers, i.e., aggrecan, type II collagen, and sox-9 expression (36, 92).

APPLICATION AND EXPOSURE OF CELLS TO DYNAMIC TENSILE STRAIN (DTF)

Twenty four hour prior to initiating experiments, TCM of the 75% to 80% confluent

monolayers of chondrocytes was replaced with TCM containing 1% FCS as described

earlier (36). Initially, cells were subjected to DTF of various magnitudes and frequencies

in the presence or absence of recombinant human interleukin-β (IL-1β). DTF at a magni-

tude of 3% and a frequency of 0.25 Hz, suppressed greater than 90% of the rhIL-1β- 43

induced iNOS mRNA expression and NO production, reproducibly. Concurrently, cells

were subjected to (A) no treatment (control), (B) rHuIL-1β (1ng/ml; Calbiochem, CA),

and D were subjected to DTF at the start of the experiment, and rHuIL-1β was immedi-

ately added to groups B and D. Cells were subjected to DTF for various time intervals,

harvested, and analyzed as required.

REVERSE TRANSCRIPTION POLYMERASE CHAIN REACTION (RT/PCR)

RNA was isolated from cells using the RNeasy kit (Qiagen, CA) after shredding through a Qiashredder (Qiagen, CA). The RNA was subjected to DNAse digestion, and stored in

40μl of RNAse-free water. The concentration and purity of RNA was spectrophotometri- cally assessed at 260 nm. The abundance of mRNA encoding for Aggrecan was examined as described earlier (36, 38). A total of 1 μg of RNA was mixed with 1μg oligo-dT

(12-18 oligomers) in RT buffer and incubated for 10 min at room temperature. This mix-

ture was transcribed with 200 units of MULV reverse transcriptase for 30 minutes at 37

°C. The cDNA was amplified with 0.1 μg of specific primers in a reaction mixture con-

taining 200 μM dNTP and 0.1 units of Taq polymerase in PCR buffer. The 3 steps of

PCR-denaturation, amplification, and extension were done in a thermal cycler (Eppen-

dorf) for 30 cycles of 40 seconds at 94 °C, 40 seconds at 62°C and 60 second at 72°C.

Each sample that underwent RT was assessed for glyceraldehyde phosphate dehydro-

genase as a standard. The sense and antisense sequence of rat primers used were as fol-

lows: GAPDH (323 bp) Sense 5’- AGACAGCCGCATCTTCTTGT-3’, antisense 5’-

TACTCAGCACCAGCATCACC-3’; Aggrecan (179 bp) Sense 5’-

CTACGACGCCATCTGCTACA-3’; antisense 5’- GCTTTGCAGTGAGGATCACA-3’.

PCR products were separated on a 2%-agarose gel at 100V/cm in Tris-acetate/EDTA electrophoresis buffer. The gels were stained with ethidium bromide and then photo- graphed in Kodak Image Station 440 CF. The intensity of the bands was analyzed by

IMAGE J (NIH, MD) program.

REAL TIME REVERSE TRANSCRITASE/POLYMERASE CHAIN REACTION

Gene specific primer sequences were selected using the Taqman Probe and Primer De-

sign function of the Primer Express v1.5 software (Applied Biosystems, CA). The sense

and antisense sequences of rat primers used were as follows: MMP-13 sense 5’-

GTTCAAGGAATCCAGTCTCTCTATGG-3’, antisense5’-

TGGGTCACACTTCTCTGGTGTTT-

3’, probe 6-FAMd (CCAAGGAGATGAAGACCCCAACCCTAAGC)BHQ-1

(XM343345); iNOS sense 5’-TTCTGTGCTAATGCGGAAGGT-3’, antisense 5’-

GCTTCCGACTTTCCTGTCTCA-3’, probe 6-FAMd

(CCGCGTCAGAGCCACAGTCCT)

BHQ-1(D44591); Cox-2 sense 5’-CTTTGGCAGGCTGGATTTTAA-3’, antisense 5’-AGAAGCCCACTGATACCTTTTGC-3’, probe 6-FAMd

(TGCACAGTATGACACAACAGCCCATCTCTC)

BHQ-1 (AF233596); MMP-9 sense 5’-AGCGCCAGCCGACTTATGT-

3’, antisense 5’-ACACAGCTGGCAGAGGATTACC-3’, probe 6-

FAMd (TCTTCCCCCAGACCTGAAAACCTCC)BHQ-1(NM_031055).

Reverse transcription reactions were carried out using 2 μg RNA and TaqMan Reverse

Transcription reagents, followed by real-time PCR using Taqman® PCR Master Mix and

ABI Prism 7700 Sequence Detection System® (Applied Biosystems, CA). Reactions

were performed as follows: Cycle I (1X): 95º C for 3.0 min, Cycle II (50X): Step 1 at 95º

C for 0.3 min, followed by Step 2 at 55º C for 0.3 min, and Step 3 at 72º C for 30 mins,

Cycle III at 40º C hold.

Following amplification, a melting curve was obtained to ensure that primer-dimers or

non-specific products had been eliminated or minimized. The data, obtained by real-time

PCR, was analyzed by the comparative threshold cycle (CT) method. In this method, the

amount of the target, normalized to GAPDH, and relative to a calibrator (either untreated sample or IL-1®-stimulated cells), is given by 2-ΔΔCT, where ΔΔCT = ΔCT (sample) - ΔCT

(calibrator), and ΔCT is the CT of the target gene subtracted from the CT of GAPDH [27].

CELLULAR PROTEIN ANALYSIS

After exposure to various treatment regimens, Flexcell membranes from each well were-

removed and cut into 4 to 6 pie shaped pieces. The cells growing on the Bioflex membranes directly over the loading posts of the Flexcell plate were analyzed, while the mar- ginal area of the well stretched by vacuum was excluded. The cellular expression of proteins was analyzed by immunofluorescence staining, using mouse anti-iNOS IgG (BD

Bioscience 610431) and FITC conjugated donkey anti-mouse IgG (Jackson

Lab715095151); rabbit anti-COX-2 (Cayman Chemicals, MI) and goat anti-rabbit-CY3; goat anti-MMP-9 (Santa Cruz, CA) IgG and donkey anti-goat-FITC (Jackson Lab, MN).

Subsequently, cells were mounted on the membrane with Vectashield (Vector Labs, CA), and observed under an epifluoresence microscope (Ziess Axioimage) or by laser scanning cytometry (LSC). At least 3 membranes from separate experiments were analyzed. On each membrane, 8 to 10 areas of 50 cells each were counted to assess the number of fluorescence positive cells by LSC or intensity of fluorescence by Zeiss Axiovision software

(Carl Zeiss Inc, Germany). For glycoseaminoglycan (GAG) analysis sections of the Flex- cell membranes exposed to various treatment regimens were fixed with ice cold methanol and stained with 0.1% Safranin-O for 10 min, and washed gently with water. Subse- quently, synthesis of GAGs in chondrocytes and pericellular area was examined micro- scopically and analyzed semiquantitatively using the Zeiss Axiovision software (Carl

Zeiss Inc, Germany). The Field Density Means were obtained from 5 different areas of each section of the membrane containing 80 to 100 chondrocytes. The mean values were calculated and presented as Field Density means per 100 cells to obtain a comparative value for GAG production in the chondrocytes following various treatments.

WESTERN BLOT ANALYSIS

For semiquantitative measurements of proteins synthesis Western blot analysis was used as described earlier (38, 40). Briefly, cells were lysed in ice cold Tris buffered saline

(TBS) containing protease inhibitor cocktail (Roche, IN), and the extracted proteins were

loaded on the SDS-10% acrylamide gels. The proteins were electrophoretically transferred to nitrocellulose membranes (Bio-Rad, CA) and identified by monoclonal mouse anti-MMP-13 IgG (1:1000 dilution; Calbiochem, CA). Monoclonal mouse anti-β-actin

IgG (1:20,000; Abcam, MA) was used to reprobe the same blots to equilibrate protein input in all lanes. Horseradish peroxidase-labeled (HRP) donkey anti-mouse IgG

(1:10,000 dilution; Chemicon, CA) or HRP-labeled goat anti-rabbit antibody (1:10,000 dilution; Santa Cruz, CA) was used as a second antibody. The presence of HRP was detected by Luminol (Amersham, IL), and the semiquantitative analysis of luminescent bands was carried out with Kodak Image Station 1000®, and Kodak 1D image analysis software.

DATA ANALYSIS AND STATISTICS

The SPSS 13.0 software (SPSS Inc., Chicago, IL) was used for statistical analysis. Each experiment was performed at least three times. For quantitative analysis, means ± S.E.M. were calculated. To determine whether significant differences exist between groups,

Oneway ANOVA and the post-hoc multiple comparison Tukey test were applied. To identify differences between IL-1β treated cells in the absence or presence of CTS at various magnitudes and frequencies, One-Way ANOVA and the post-hoc multiple comparison Dunnett test were used. Differences were regarded as statistically significant at values of P<0.05

Results

INHIBITION OF IL-1β DEPENDENT INOS INDUCTION BY DTF IS LONG-LASTING

Previous reports provided evidence that mechanical signals of low magnitudes inhibit the

IL- 1β-dependent proinflammatory gene transcription [13, 22]. However, these reports did not address how long the effects of these signals persist in an inflammatory environment following removal of biomechanical stimulation. Furthermore, the duration of biomechanical stimulation that is essential for sustained anti-inflammatory effects remains unknown. To gain insight into these questions, we first determined IL-1β-induced iNOS mRNA expression in chondrocytes continuously exposed to DTF for various time intervals. As shown in Fig 4 A inset, IL-1β induced a marked upregulation of iNOS mRNA in chondrocytes. Simultaneous exposure of cells to DTF for 4, 8, or 16 h blocked more than

90% of IL-1β- induced iNOS expression at all time points tested. To further determine how long the signals generated by DTF persisted, chondrocytes after addition of IL-

1β were immediately subjected to DTF, after various time intervals DTF was removed and cells allowed to rest in IL-1β containing medium (DTF/rest) for additional periods of time, and analyzed for IL-1β- induced iNOS expression. Fig. 4 A demonstrates that, cells exposed to DTF/rest for 4/20, 8/16, 12/12, 16/8, 20/4, or 24/0 h, suppressed IL-1β- induced iNOS expression by 54%, 93%, 90%, 84%, 80%, and 90%, respectively, despite the presence of IL-1β. This suggested that a 4 h exposure to DTF, that inhibits 97% of IL-

1β-induced iNOS expression (inset Fig. 4 A), was not sufficient to inhibit iNOS expression 20 h later. Nevertheless, maximal inhibition of iNOS expression was observed following an 8 h of DTF exposure which persisted for 16 h (Fig. 4A). LSC analysis of chondrocytes immunostained for iNOS revealed that the transcriptional downregulation of iNOS expression by DTF was also reflected in a significant reduction (p<0.05) in the number of iNOS positive cells (Fig. 4 B).

Specifically, cells exposed to 8, 12 or 16 h of DTF, followed by 16, 12 or 8 h of rest exhibited lowest iNOS synthesis as compared to IL-1β treated controls. To further determine how long the DTF-mediated suppression of IL-1β actions persists, we next exposed

IL-1β-treated chondrocytes to DTF for 8 h followed by a rest for the ensuing 16, 28, or

40 h. The results demonstrated DTF exposure for 8 h caused a 90% inhibition of iNOS mRNA expression 16 h later, but the same exposure failed to suppress iNOS mRNA expression 28 or 40 h later (Fig. 4 C). Concomitant measurements of NO release in the culture supernatants, measured by Griess reaction, revealed that NO production was also inhibited by 64% in chondrocytes exposed to DTF for 8 h followed by a 16 h rest

(Fig. 1 D). However, DTF exposure for 8 h was insufficient to block IL-1β-induced NO

production 28 or 40 h later. This suggested that 8 h of DTF is sufficient to inhibit

proinflammatory gene induction for a limited period of time, beyond which effects of IL-

1β become apparent again (Fig. 4 D).

SUSTAINED DOWN-REGULATION OF COX-2 INDUCTION BY DTF

To confirm that DTF regulates transcriptional activation of multiple proinflammatory genes upregulated in arthritic joints, we next determined the effects of DTF on COX-2 induction. The extent of the persistence of DTF signals on COX-2 mRNA expression was evaluated by subjecting the chondrocytes to DTF and IL-1β as described above. Quantita- tive analysis of PCR products showed a low level constitutive COX-2 mRNA expression in chondrocytes. As shown in Fig. 2 A, the signals generated by DTF/rest of 4/20, 8/16,

12/12, 16/8, 20/4, or 24/0 h, significantly inhibited COX-2 mRNA expression by 37%,

72%, 81%, 88%, and 83%, respectively (Fig. 5 A). Nevertheless, signals generated by

DTF which could inhibit 96% of IL-1β-induced COX-2 mRNA expression (inset Fig. 5

A) failed to inhibit more than 37% of COX-2 mRNA expression following a 20 h rest

(Fig. 2 A).

Quantitative analysis by LSC of cells immunostained for COX-2 revealed that DTF also inhibits its synthesis in a persistent manner in parallel to inhibition of COX-2 mRNA

(Fig. 5 B). Similarly, despite the presence of IL-1®, signals generated by DTF/rest for

4/20, 8/16, 12/12, 16/8, or 20/4 h blocked COX-2 induction by a significant 68%, 63%,

67%, 76%, or 78%, 14 respectively. Interestingly, exposure of cells to 24 h of continuous

DTF without rest revealed more than 94% suppression of COX-2 induction, indicating that constant exposure to DTF is persistent suppressor of COX-2 (Fig. 5 B and C).

As evident in Fig. 2 C, COX-2 induction by IL-1β coincided with its nuclear transloca-

tion. Not only did DTF exposure down-regulate IL-1β-induced synthesis of COX-2,

but also its nuclear translocation. Exposure of cells to DTF/rest for 8/16 or 16/8 demon-

strated a significant reduction in the presence of COX-2 in both cytoplasmic and nuclear

compartment (Fig. 5 C). Interestingly, while a complete inhibition of COX-2 mRNA

expression was not observed at any time point tested, LSC analysis failed to detect the presence of COX-2 in cytoplasmic or nuclear compartments of untreated control chondrocytes.

DTF-MEDIATED SUPPRESSION OF IL-1 β-DEPENDENT MMP INDUCTION

Chondrocytes synthesize a number of MMPs among which MMP-13 and MMP-9 are

major mediators of matrix degradation in RA and OA (16, 93). Because both of these

MMPs degrade collagenous and non-collagenous proteins in the cartilage, potential of

DTF on their IL-1β-induced induction was examined. To determine whether continuous

application of DTF results in the inhibition of IL-1β-induced MMP-9 and MMP-13

mRNA expression, chondrocytes exposed to IL-1β alone for 4 or 24 h. As shown in Fig.

6 A inset, IL-1β induced a marked upregulation of MMP-9 and MMP-13 mRNA in chon-

drocytes. Simultaneous exposure of cells to DTF for 4, or 24 h blocked more than 90% of

IL-1β-induced MMP-9 and MMP-13 expression. To further determine whether DTF elic-

its sustained effects on the inhibition of MMP-9 and MMP-13 expression, cells were sub-

jected to DTF/rest for 4/20, 8/16, 12/12, 16/8, 20/4, or 24/0 h, in the presence of IL-1β.

Subsequent mRNA analysis demonstrated that like iNOS and COX-2, DTF continued to block both MMP-9 and MMP-13 mRNA expression persistently after its removal. An 8 h or longer exposure to DTF was required for its persistent effects. Again, constant application of DTF for 4 h which inhibited more than 90% of MMP-9 and MMP-13 mRNA expression (inset Fig. 6 A), exhibited only a 52% (MMP-9) and 68% (MMP-13) inhibition following removal of DTF for 20 h, indicating that the signals generated by DTF persist for a limited time and the IL-1β-induced proinflammatory gene expression resumes after removal of DTF.

We further confirmed that DTF mediated suppression of IL-1β induction of MMP-9 and MMP-13 mRNA was reflected in sustained inhibition of their synthesis (Fig 6B).

Quantitative analysis of fluorescence by LSC in cells immunostained for MMP-9 and

Western blot analysis for MMP-13 revealed that signals generated by DTF also blocked

MMP-9 and MMP-13 synthesis in a sustained manner. As apparent in Fig. 6 B, protein synthesis in chondrocytes in response to DTF demonstrated a pattern similar to mRNA expression for both MMP-9 and MMP-13. Notably, a 4 hr of DTF exposure was insufficient to sustain the MMP-9 and MMP-13 inhibition more than 51% and 69%, respectively (Fig. 6 B). DTF/rest for 8/16 exhibited maximal inhibition of both MMP-9 and

MMP-13, 16 h later. On the contrary, examination of protein synthesis after DTF/rest for

12/12, 16/8, or 20/4 h revealed a suppression of 75%, 66%, or 77%, respectively, which significantly less than that observed by DTF/rest of 8/16 (Fig. 6 B and C).

DTF-MEDIATED UPREGULATION OF AGGRECAN SYNTHESIS IS SUSTAINED FOLLOWING

REMOVAL OF MECHANICAL SIGNALS

To be clinically effective as a therapeutic signal in joint repair, biomechanical signals must exhibit reparative actions. Since aggrecan constitutes the major protein in the

GAGs, next experiments were designed to examine the effects of DTF on aggrecan gene transcription. Chondrocytes expressed high levels of aggrecan mRNA constitutively, and this expression was markedly inhibited (95±4%) by treatment of cells with IL-1β for 4 or

24 h. Simultaneous exposure of cells to DTF and IL-1β for 4 or 24 h rescued IL-1β- induced inhibition of aggrecan gene expression by 62% or 41%, respectively (Fig. 7 B).

To determine whether the signals generated by DTF persistently upregulate aggrecan gene transcription, chondrocytes were subjected to the same regimen of DTF/rest as described above. Examination of cells exposed to DTF/rest for various time intervals, revealed that DTF consistently revoked IL-1β-dependent inhibition of aggrecan mRNA expression persistently at all time points tested, indicating an anti-catabolic role of DTF on chondrocytes. For example, a 4.57, 4.62, 5.71, 6.71, 12.28, or 9.57 fold increase in mRNA was observed in cells exposed to DTF/rest for 4/20, 8/16, 12/12, 16/8,

20/4, or 24/0 h, respectively (Fig. 7 A).

Thus, cumulatively these data indicated that DTF exerts potent anabolic effects by miti- gating the IL-1β-mediated inhibition of aggrecan transcriptional regulation. The semiquantitative analysis of GAG contents in chondrocytes revealed a marked inhibition of

total GAG contents in IL-1β treated cells, as compared to untreated control cells. Expo-

sure of DTF rescued GAG synthesis by counteracting IL-1®-induced GAG inhibition.

The effects of DTF were persistent, and could be observed following removal of DTF.

For example, chondrocytes exposed to DTF/rest for 4/20, 8/16, 12/12, 16/8, 20/4 and

24/0 h in the constant presence of IL-1β, exhibited a 245%, 252%, 287%, 330%, 370%,

or 265%, greater total GAG contents as compared to IL-1β-treated cells (Fig. 7 C).

Nevertheless, GAG synthesis was below control untreated cells unless cells were treated

with DTF/rest for 16/8 or 20/4 h.

Discussion

The present findings provide molecular basis for the persistent actions of biomechanical signals in chondrocytes. While earlier studies demonstrated that signals generated by both tensile and compressive forces of low magnitudes inhibit the IL-1β- dependent proinflammatory gene transcription (18, 20, 33, 36, 38, 40), the persistence of these signals in a proinflammatory environment is as yet unclear. Studies directed to- wards arthritic joint healing in vivo, have demonstrated that controlled application of biomechanical stimulation reduces inflammation and enhances wound healing in inflamed or injured cartilage (61. 79, 80). However, these reports do not address the persistence of biomechanical signals in relation to the duration of motion-based therapies. To answer such key questions, we have used primary cultures of chondrocytes to examine

the long term effectiveness of biomechanical strain in attenuating inflammatory mediator production and augmenting the extracellular matrix synthesis in vitro. We show that signals generated by DTF are sustained for prolonged periods of time following the removal of mechanical stimulation, and continue to act as potent anti-inflammatory signals in a proinflammatory environment.

The examination of the sustained effectiveness of DTF as an anti-inflammatory signal is important to the application of biomechanical therapy in inflamed cartilage. Therefore, to gain broader understanding of the role of biomechanical signals in the preservation of cartilage, we have focused on the regulation of 4 major proinflammatory mediators with distinct functions in cartilage destruction (43, 61, 74, 81, 94, 95), namely iNOS, COX-2,

MMP-9 and MMP-13. All of these mediators are produced by chondrocytes and are ele- vated in arthritic joints. Inducible NOS and COX-2 are involved in the amplification of proinflammatory responses. NO is cytotoxic to chondrocytes, damages cartilage, and is involved in the upregulation of MMPs (94, 96, 97). COX-2 induces prostaglandins that in turn upregulate production of inflammatory mediators including cytokines to initiate cartilage destruction (95, 98). MMP-9 and MMP-13 are directly responsible for proteolytic degradation of cartilage. In addition to collagen, MMP-13 also degrades the proteoglycans including aggrecan, giving it a dual role in matrix destruction. MMP-9 is involved in degradation of non-collagen matrix components of the joints (61, 84).

The present findings demonstrate a number of important points with regard to the biomechanical modulation of articular chondrocyte phenotype and functions. First,

continuous exposure of chondrocytes to DTF induces a marked suppression of IL-1β-

dependent proinflammatory gene induction, during the entire duration of exposure to

DTF of appropriate magnitude and frequency. However, these studies were limited to a

24 h window. Whether the effectiveness of biomechanical signals is reduced after longer durations is yet to be elucidated. Additionally, exposure of chondrocytes to continuous

DTF from 4 to 24 h inhibits IL-1β-induced iNOS, COX-2, MMP-9, and MMP19 13, mRNA expression and synthesis, suggesting that the effects of DTF are not limited to one or two genes, rather it affects are broad and modulate expression of several proinflammatory genes.

The second important point is that the effects of DTF are persistent and continue to attenuate IL-1β-induced proinflammatory gene transcription for hours after the removal of DTF even in an inflammatory environment. DTF-mediated modulation of proinflammatory gene expression is long lasting, i.e., if chondrocytes are exposed to DTF for 8 h, the signals generated by DTF continue to block iNOS gene transcription and synthesis for the next 16 h.

Furthermore, extending the time of exposure to DTF followed by a shorter interval of rest does not significantly alter the persistence of signals generated by DTF. Additionally, after removal of DTF the signals continue to suppress multiple proinflammatory genes like

COX-2, MMP-9 and MMP-13 in a persistent manner, despite the presence of a proinflammatory environment. Both of the above observations suggest that the suppression of

proinflammatory genes occurs at the transcriptional level. Previously it has been shown that DTF inhibits IL-1β-induced proinflammatory genes via suppression of nuclear translocation of NF-κB [21]. This may explain the DTF-mediated inhibition of multiple proinflammatory genes that are under the control of NF-κB [38]. It is also likely that the persistent effects of DTF in inhibiting proinflammatory gene induction are also controlled by

NF-k⎢B transcription factors. The third significant finding is that, during biomechanical modulation of chondrocyte functions, the duration of exposure to DTF is a critical deter- minant for the length of its persistent effects on cells. This is evidenced by the observations that constant DTF exposure 20 for 4 h leads to greater than 90% inhibition of iNOS mRNA expression, whereas, a 4 h DTF exposure followed by a rest of 20 h results in the re-expression of all proinflammatory genes examined, i.e., iNOS, COX-2, MMP-9 and

MMP-13. Similarly, constant DTF exposure for 8 h blocks more than 95% of iNOS gene transcription, and the level of inhibition of iNOS mRNA expression and synthesis is maintained following a 16 h of rest. Nevertheless, 8 h exposure of DTF was insufficient to block iNOS gene transcription and NO production 28 or 40 h later. These findings suggest that the extent of persistence of anti-inflammatory signals generated by DTF is dependent upon the duration of exposure of DTF to the chondrocytes.Additionally, DTF blocks expression of proinflammatory proteins transiently and these genes re-express following a defined period of rest in a proinflammatory environment. It is important to note that DTF alone as a constant signal, or after a period of rest, does not induce proinflammatory signals in chondrocytes (data not shown data is shown in 24/0 treatments). In this regard, the magnitudes of mechanical signals play an important role in inducing or attenuating proinflammatory signals (15, 16, 18, 20, 21, 33, 38, 40, 43, 83, 89, 90, 91). 58

LSC analysis in our studies revealed that IL-1β-induces iNOS production in all

chondrocytes. However, the re-expression of iNOS after removal of DTF, varied among

chondrocytes. In cartilage, chondrocytes are shown to be of different types that differ in

their responses to IL-1β (99). Therefore it is feasible that DTF removal leads to iNOS in-

duction in a certain population of cells in the presence of a proinflammatory stimulus. On

the contrary, LSC analysis demonstrated that DTF blocked expression of COX-2, MMP-

9 and MMP-13 to the same degree in all cells. Furthermore, re-expression of these pro-

teins following removal of DTF was also similar in all cells. These findings again point to

the fact that the sustained effects of DTF in blocking proinflammatory gene expression

may be controlled at transcriptional level.

Finally, during biomechanical modulation of chondrocytes, signals generated by DTF

are not only effective in suppressing proinflammatory gene induction, but also in un-

blocking the IL-1β-induced inhibition of aggrecan synthesis. The primary mechanisms

for IL-1β-mediated cartilage destruction involve activation of proinflammatory genes on

one side, and inhibition of matrix synthesis on the other side (36, 79). The above findings

suggest that DTF not only limits the inflammation, but also augments repair of cartilage.

For example, IL-1β-induced blockage of aggrecan mRNA expression is rapidly abrogated by DTF within 4 h, and these signals persist following removal of DTF for ensuing 20 h.

Thus, biomechanical signals modulate proinflammatory genes and aggrecan transcription

and synthesis in a diametrically opposite manner. These findings are of interest, because

aggrecan gene expression is controlled by AP-1 transcription factors (100), and suggest

that signals generated by DTF may simultaneously act on more than one signaling cas-

cades to control inflammation and initiate repair of inflamed chondrocytes.

In conclusion, signals generated by DTF are converted into potent biochemical events

that block the synthesis of catabolic mediators and induce anabolic effects in a persistent

manner, long after the removal of biomechanical stimuli. Within the constraints of these

in vitro observations, the clear evidence points to the fact that biomechanical forces may

indeed induce sustained anabolic responses in inflamed cartilage. Our findings serve as a

foundation for in vivo studies, and demonstrate the necessity of using adequate time in-

tervals for sustained effects of motion-based therapies in the optimal management of

acute /chronic inflammation of the joints.

Acknowledgements

This research was supported by NIH grant numbers AR 04878, AT00646 & DE015399.

Fig. 4. Effect of DTF on iNOS induction in articular chondrocytes. Chondrocytes grown on Bioflex plates were exposed to DTF for 0, 4, 8 or 16 h, in the presence of IL-1β (inset

A), or DTF followed by rest (DTF/rest) for 4/20, 8/16, 12/12, 16/8, 20/4 or 24/0 h, in the constant presence of IL-1β (A). The expression of iNOS mRNA was analyzed by Real

Time PCR (A, and inset A), and iNOS protein was assessed in immunostained chondrocytes by Laser Scanning Cytometry (B). Analysis of iNOS mRNA expression in chondrocytes exposed to DTF/Rest for 8/16, 8/28, or 8/40 by Real Time PCR (C), and total

NO accumulation in the culture supernatants of chondrocytes exposed to DTF/Rest for

8/16, 8/28, or 8/40 (D). Data represent mean and SEM of three separate experiments performed in triplicates. *indicates p<0.05 as compared to IL-1β treated cells.

Figure 4

Fig. 5. Effect of DTF on IL-1β-dependent Cox-2 induction in articular chondrocytes.

Chondrocytes grown on Bioflex plates, were exposed to DTF for 4 or 24 h, in the presence of IL-1β (inset A), or DTF/rest for 4/20, 8/16, 12/12, 16/8, 20/4 or 24/0 h, in the constant presence of IL-1β (A). The expression of COX-2 mRNA was analyzed by Real

Time PCR (A, and inset A), and COX-2 protein was assessed in immunostained chondrocytes by LSC (B). Microscopic representation of chondrocytes treated with DTF/rest for 0/24, 8/16, 16/8, or 24/0 h in the presence of IL-1β showing nuclear localization of

COX-2. Chondrocytes were stained with anti-COX-2 antibodies to show the sustained inhibition of COX-2 synthesis in cells treated with DTF/rest for various time intervals

(C). Data in (A and B) represent mean and SEM of three separate experiments performed in triplicates. Micrographs in (C) represent one of three separate experiments. *indicates p<0.05 as compared to IL-1® treated cells.

Figure 5

Fig. 6. DTF-induced blocking of IL-1β-dependent induction of MMP-9 and MMP-13.

Articular Chondrocytes grown on flexible bottom plates were exposed to various durations of DTF or DTF/rest. Chondrocytes exposed to DTF for 4 and 24 hrs in the presence of IL-1β and analyzed for MMP-9 and MMP-13 by Real Time PCR (Inset Fig 6 A).

Chondrocytes exposed to DTF/rest for 4/20, 8/16, 12/12, 16/8, 20/4 or 24/0 h were analyzed by Real time PCR for MMP-9 and MMP-13 mRNA expression (A). Synthesis of

MMP-9 was assessed by LSC evaluation of Immunostained cells (B and C) and synthesis of MMP-13 was determined by densitometric analysis of Western Blots (B). Data represent mean and SEM of three separate experiments. * indicates p<0.05 as compared to IL-

1β treated cells. indicates p<0.01 as compared to chondrocytes treated with DTF/rest for 8/16 h.

Figure 6.

Fig. 7. Effect of DTF on IL-1β induced attenuation of aggrecan synthesis. Chondrocytes cultured on Bioflex plates were subjected to DTF or DTF/rest in the presence of IL-1β.

Cells exposed to DTF/rest for 4/20, 8/16, 12/12, 16/8, 20/4, or 24/0, were analyzed for

GAPDH and aggrecan mRNA expression by RT-PCR (inset A), and the intensity of each

PCR product measured by densitometric analysis (A). RT- PCR analysis of chondrocytes exposed to DTF for 0, 4, or 24, h, in the presence of IL-1β (B). GAG synthesis in chondrocytes exposed to various periods of DTF/rest as shown in (A), were examined by Sa- franin-O staining, using dichromatic filter and Axioplan Software. The bars represent means and SEM of Field Density Mean per 100 cells in 5 different areas of membrane, ∗ represents p<0.05 as compared to IL-1β treated cells (C)

Figure7.

CHAPTER 4

BIOMECHANICAL SIGNALS SUPPRESS TAK1

ACTIVATION TO INHIBIT PROINFLAMMATORY GENE

INDUCTION IN FIBROCHONDROCYTES

INTRODUCTION

Arthritic diseases are chronic inflammatory diseases of the joints associated with signifi-

cant cartilage and bone erosion resulting in compromised joint function. Increased pro-

duction of cytokines including interleukin-1 (IL-1) and tumor necrosis factor-α (TNF-α)

(101, 102) by the synoviocytes and chondrocytes provide evidence for their involvement in the pathogenesis of both, rheumatoid (RA) and osteoarthritis (OA). These cytokines upregulate transcription of proinflammatory genes such as inducible nitric oxide synthase

(NOS2A), cyclooxygenase-2 (COX-2), matrix metalloproteinases (MMPs), IL-1β and

TNF-α to initiate cartilage destruction and amplify immune responses (103, 104). Tar- geted therapeutic strategies like neutralization of TNF-α or IL-1 activity by antibodies, or

delivery of IL-1 receptor antagonist (IL-1Ra) reduces the severity of the disease in ex-

perimental and human RA (105). However, anti-inflammatory drugs are the choice treat-

ment, whereas inhibitors of NF-κB or gene silencing of IKKβ are shown to suppress cartilage and bone erosion in experimental models of arthritis (98, 106-108). Thus far most of these therapies have met with limited success, and arthritis remains one of the most common ailments and major causes of morbidity in the US population.

While the therapeutic potential of mobilization in restoring joint function in ar-

thritic diseases is well recognized for centuries, relatively little is understood about the

intracellular events underlying the actions of biomechanical signals in inflamed cartilage.

These signals are essential for cartilage homeostasis, as evidenced by the findings that

immobilization of healthy joints results in cartilage matrix loss (109). Chondrocytes em-

bedded in the cartilage matrix constantly experience compressive, tensile, and shear forces during joint mobilization (112). These cells are mechanosensitive and can perceive and respond to biomechanical signals transduced by cell surface molecules such as β-

integrins and focal adhesion complexes (110-114). At high or traumatic magnitudes bio-

mechnical signals trigger expression of proinflammatory genes in chondrocytes, and at

low physiological magnitudes these signals are potent inhibitors of IL-1β- and TNF-α-

dependent proinflammatory gene transcription (37, 38, 49). Furthermore, at low magni-

tudes these signals induce proteoglycan and collagen type II synthesis essential for cartilage homeostasis and repair (33, 36- 38, 49, 115, 116).

TGF-beta activating kinase, TAK1, is a key regulator of the inducible transcription factor NF-κB, and thus is pivotal in regulating down stream signaling events that mediate proinflammatory gene transcription. In Tak1m/m cells, signaling initiated by

TNF/IL-1 and Toll like receptors (TLR)-3 and 4 is severely impaired. However, cells

lacking TAK1 associated binding proteins (TAB)-1 and -2 exhibit normal NF-κB activa-

tion, indicating that signals generated by all three major NF-κB activators converge at

TAK1 (116). TAK1 in turn activates the signalosome, a complex comprising of three

distinct I-kB kinases (IKK), IKKα, ΙΚΚβ and IKKγ. In canonical pathway, upon phos-

phorylation by IKKβ, ΙκΒα and IκBβ, the cytoplasmic inhibitors of NF-κB, are marked

for ubiquitination and subsequent proteosomal degradation. The degradation of IkBα and

IkBβ allows phosphorylation of NF-κB at multiple sites in a stimulant dependent manner

and its transmigration to the nucleus. The binding of NF-κB to its consensus sequences

leads to transcription of a plethora of genes including proinflammatory cytokines and

mediators, as well as several of the molecules required for the activation of NF-κB sig-

naling cascade. Although this classical model of NF-κB activation by TNF-α or IL-1β is

well documented, its complexity evolves from its regulation at multiple intracellular lev-

els, in a cell as well as stimulus dependent manner.

Since biomechanical signals inhibit proinflammatory gene transcription, we examined whether these signals regulate specific steps in the NF-κB pathway to attenuate the proinflammatory responses. Fibrochondrocytes (FCs) from the fibrocartilagenous disk of the temporomandibular joint (TMJ) were isolated by enzymatic digestion (40), and exposed to cyclic tensile strain (CTS) to mimic tensile forces perceived by cells during TMJ mobilization (40, 49). To investigate the mechanisms of anti-inflammatory actions of biomechanical signals with in the NF-κB signaling cascade, cells were exposed to IL-1β as a proinflammatory signal, CTS alone, or simultaneously to CTS and IL-1β, and compared to control untreated cells. Following exposure to CTS and/or IL-1β for various time intervals, cells were harvested to examine the mRNA expression, regulation of proteins and their nuclear and cytoplasmic localization, within the NF-κB pathway.

Materials and Methods

Fibro-chondrocyte harvesting and cell-culture

Articular discs of the temporomandibualr joints of 10-12-week-old female Sprague-

Dawley rats (Harlan, IN), were harvested following approval by the Institutional Labora- tory Animal Care and Use Committee at The Ohio State University. Cartilage pieces were minced in Hank’s balanced salt solution (HBSS; Invitrogen, CA), transferred onto a macroporous filter (Spectra/Mesh, CA) and placed in a digestion chamber. After treatments for 10 min with 0.2% trypsin and 2 h with 0.2% collagenase type II (Worthington,

NJ), cells released from the cartilage were pelleted at 800 x g for 5 min, and washed

twice with HBSS. The cells were then cultured in TCM [(DMEM/F12 1:1) (Mediatech,

VA), 10% FBS (Hyclone, UT), 10 µg/ml penicillin and 100 U/ml streptomycin (Medi-

atech, VA), and 2 mM L-glutamine (Invitrogen, NY)]. Fibrochondrocytes were used dur-

ing the first 3 passages. Between these passages cell morphology and the mRNA expres-

sion for phenotypic markers, aggrecan, biglycan and versican were stable (36).

Application of Cyclic Tensile Strain (CTS)

Cells (5 X 104/well) were grown for 3-4 days on collagen I-coated BioFlex 6-well culture plates (Flexcell Int, NC) to 80% confluence in 5% CO2 and 37ºC. Twenty four h before

start of the experiments, the medium was replaced with TCM with 1% FBS, and the

plates were placed onto a Flexcell loading station and subjected to CTS at a magnitude of

12% and frequency of 0.05 Hz. Four different treatment regimens were assigned: i) un-

treated controls, ii) cells treated with recombinant human IL-1β (1 ng/ml; rhIL-1β; Cal-

biochem, CA), iii) cells treated with CTS, iv) cells treated with CTS and rhIL-1β.

Analysis of mRNAs expression by real-time PCR

RNA was isolated using the RNeasy Mini Kit as recommended by the manufacturer

(Quiagen, CA). RNA (1µg) was reverse transcribed into cDNA in a total volume of 30 µl

using 200 U of Moloney murine leukemia virus, first at 42ºC for 25 min and then at 65ºC

for 5 min. Real time PCR was performed in a volume of 25 µl on a Biorad iCycler iQ

(Biorad, CA) as described previously (39). Following primers and probes were used: iN-

OS-sense 5’- TTCTGTGCTAATGCGGAAGGT -3’, iNOS-anti-sense 5’-

GCTTCCGACTTTCCTGTCTCA-3’, iNOS-probe 5’-

CCGCGTCAGAGCCACAGTCCT-3’; I-κBα-sense 5’-

GGTATACTTAGCACCACAGCACACA -3’, I-κBα-anti-sense 5’-

CCCCAAATTTCACAAGAACAACA -3’, I-κBα-alpha-probe 5’-CCT

AGCCCCGAGCATTCTATTGTGGTGAT-3’; I-κBβ-sense 5’-CCATGTAGC

TGTCATCCACAA AG-3’, I-κBβ-anti-sense 5’-ACG TAG GCT CCG GTT TAT GAG

-3’, I-κBβ-probe 5’-AGAGATGGTCCAACTGCTCAGGGATGCT -3’, MMP-13-sense

5’- GTT CAA GGA ATC CAG TCT CTC TAT GG-3’ ; MMP-13-anti-sense 5’ - TGG

GTC ACA CTT CTC TGG TGT TT- 3’ ; MMP-13-probe 5’ - CCA GGA GAA GAC

CCC AAC CCT AAG C- 3’.

Protein phosphorylation and degradation

Activation and synthesis of transcription factors was analyzed by Western blot analysis.

Total cell extracts were separated by SDS-10% polyacrylamide gel electrophoresis (SDS-

PAGE) and blotted on PVDF membranes (BioRad, CA). The blots were probed with anti-IκB-α, anti-IκBβ (Santa Cruz Biotechnology, CA), and phospho-IκBα Serine32, and

Serine36, phospho-NF-κB p65 Serine536, and Serine276, anti-NF-κB p65, anti-TAK1 and phospho-TAK1 Threonine187 (Cell Signaling, MA) antibodies and the binding of

primary antibodies revealed with IR-Dye conjugated secondary antibodies (Licor Biosci-

ences, NE). The Blots were imaged with the Odyssey Infrared Imaging System and soft-

ware (Licor Biosciences, NE) at 700 and 800 nm channels and 169 µm resolution..

Estimation of I-κB kinase-α (IKK-α) Activation

IKK activation in response to IL-1β and CTS was investigated as was described earlier

(117). Briefly, cells were suspended in lysis buffer (LB; 20 mM Tris (pH 8.0), 0.5 M

NaCl, 0.25% Triton X-100, 1 mM EDTA, 1mM EGTA, 10 mM β-Glycerolphosphate),

for 20min at 4ºC. IKK complexes were immunoprecipitated with anti-IKKγ antibodies

(Santa Cruz Biotech, CA), washed with LB and then incubated with 2 μg of GST-IκBα

conjugated to agarose beads (Santa Cruz, CA) and 0.5 µmol of cold ATP at 30ºC for 2 h.

The beads were thoroughly washed with kinase buffer and the GST-IκBα was size sepa-

rated by SDS-10%PAGE, transferred to nitrocellulose membranes and probed with either

anti-phospho Ser 23/25-IκBα (Cell Signaling, MA) or anti-GST antibodies (Pharmacia,

NJ). IRdye 800 goat anti-rabbit antibody or IRdye 680 goat anti-mouse antibodies were

used as a second antibody. The Blots were imaged with the Odyssey Infrared Imaging

System and software (Licor Biosciences, NE) at 700 and 800 nm channels and 169 µm resolution.

Immunofluorescence

Activation and synthesis of NF-κB, I-κBα and I-κBβ was assessed by immunofluorescence using rabbit anti-NF-κB p65 IgG, anti-IκB-α, anti-IκB-β or mouse anti-IκBα monoclonal antibodies for double staining (Santa Cruz Biotech, CA) as primary antibodies and CY2 or Cy3-conjugated goat anti-rabbit IgG as secondary antibody (Jackson Labs,

ME). The stained cells on Bioflex membranes were mounted and visualized under a Zeiss

Axioimage epifluorescence microscope and densitometrically analyzed using Zeiss Ax- iovision software (Zeiss, Germany).

Electrophoresis mobility shift assay (EMSA)

Native polyacrylamide gels (5 %) were prepared in Tris-borate-EDTA buffer. Fifty na- nomoles of IRDye-labeled NF-κB consensus oligonucleotides (5’-

AGTTGAGGGGACTTTCCCAGGC-3’; 3’-TCAACT CCCCTGAAAGGGTCCG -5’

Licor Biosciences, NE) were added to the EMSA binding buffer (Lightshift EMSA kit;

Pierce, IL). A total of 20 µg of nuclear proteins from cells exposed to various treatment regimens were added to the mixture, and incubated at room temperature for 20 min. The protein-DNA complexes were separated from unbound DNA by electrophoresis through a 5 % nondenaturing polyacrylamide gel at 100 V for 1 h in a 0.5 x TBE buffer. Subse- quently the gel was transferred to and imaged on an Odyssey Infrared imaging system

(Licor Biosciences, NE)

Immunoprecipitation of IκB-α and NF-κB P65

Primary FCs were treated with ice-cold RIPA buffer (50 mM Tris-HCl, pH 7.4, 150 mM

NaCl, 0.1% SDS, 1% NP-40, 1% deoxycholic acid sodium) containing a cocktail of pro-

tease inhibitors (Roche, NJ). The pre-cleared nuclear lysates were incubated with 2 mg

mouse anti-IκBα antibody and 20 ml protein A/G Sepharose beads at 4º C overnight with

end-over-end rotation. The precipitated immune complexes were separated on 10% SDS-

PAGE, electrophoretically transferred to nitrocellulose membranes, incubated with either rabbit anti-Nf-κB or rabbit anti-IκBα (diluted 1:1000) and the corresponding IRDye-

conjugated secondary antibodies, visualized with Licor Odyssey

Statistical analysis

All data were analyzed by SPSS 13.0 software (SPSS Inc., Chicago, IL). At least three independent experiments were performed and the most representative data has been presented. One-Way ANOVA and the post hoc multiple comparison Dunnett’s test were used to for statistical analysis (* = p ≤ 0.05) and comparisons were made between

stretched IL-1β-treated cells and unstretched IL-1β-treated cells.

Results

CTS attenuates TNF-α, IL-1β and LPS induced NOS2A and MMP-13 induction

To explore the mechanisms of actions of biomechanical signals, we first investigated whether these signals inhibit proinflammatory gene induction by a variety of inflammatory mediators. In these experiments, FCs were exposed to CTS of predetermined low magnitudes (12% tension at 0.05 Hz) known to inhibit IL-1β-induced gene transcription

(40, 49). Stimulation of FCs with three major proinflammatory mediators IL-1β, TNF-α

or LPS, individually or in combination, resulted in a significant increase in NOS2A and

MMP-13 mRNA expression with in 3 h. Consistent with the antiinflammatory actions of biomechanical signals, CTS suppressed more than 70% of the NOS2 and MMP-13 expression induced by IL-1β, TNF-α and LPS alone, and in combination with IL-1β and

TNF-α, or IL-1β, TNF-α and LPS (Fig 8 A and B). NOS2A expression was not observed in untreated control cells or cells treated with CTS alone. This suggests that CTS at these magnitudes acts in an inflammatory signal dependent manner to attenuate proinflammatory gene induction.

Because, the above studies demonstrated that CTS suppressed proinflammatory gene induction induced by all the 3 inflammatory mediators, IL-1β, TNF-α and LPS, re- quiring distinct receptors for cell activation, we next investigated whether the downregulation of NOS2 expression by CTS was mediated via disruption of receptors for

TNF-α, IL-1β and/or LPS. In order to explore whether CTS disrupts receptors for IL-1β,

TNF-α or LPS, cells were pre-exposed to CTS for 1 h, prior to addition of IL-1β, TNF-α, or LPS to cells and the extent of NOS2A induction was examined 15 h later. Cells were also exposed to IL-1β, TNF-α, or LPS in the presence or absence of CTS. In response to all three stimuli, FCs produced significantly robust levels of NOS2 (Fig 8 C). Cells pre- exposed to CTS for 1 hr, prior to addition of IL-1β, TNF-α or LPS demonstrated NOS2A production that was similar to cells that were not pre-exposed to CTS, indicating that preexposure of cells to CTS did not disrupt their receptors to impair their ability to respond to IL-1β, TNF-α or LPS. As expected cells exposed to CTS for 1 h and then exposed simultaneously to CTS and IL-1β, TNF-α or LPS, exhibited attenuation of NOS2 gene expression further providing evidence that pre-exposure of cells to CTS neither compro- mises their ability to respond to IL-1β, TNF-α or LPS or the ability of CTS to suppress

IL-1β, TNF-α or LPS-induced NOS2 expression (Fig 8 C).

CTS inhibits IL-1β, TNF-α and LPS-dependent TAK1 phosphorylation

Above experiments suggested that CTS intercepts a common target utilized by all three

proinflammatory stimuli to inhibit activation of NF-κB transcriptional activity. Signals

generated by IL-1β, TNF-α, or LPS receptors converge at TAK1 to trigger its rapid acti-

vation while TAB1 and TAB2 are dispensable for activation of NF-kB (116). Therefore,

we examined the CTS-mediated regulation of TAK1 activation, but not TAB1 and TAB2.

CTS inhibited IL-1β, TNF-α, or LPS-induced TAK1 phosphorylation within the first 10 min by 73%, 84%, and 56%, respectively (Fig 9 A).

CTS inhibits IL-1β-dependent IKK activation

IKKβ, the principal kinase in the classical NF-κB pathway phosphorylates IκB proteins marking them for degradation. Since IKKα and -β are tightly bound to NEMO(IKKγ), through their NEMO-binding domain, IKK complexes from cells subjected to IL-

1β alone or IL-1β and CTS for 10, 30, 60 and 90 min, were immunoprecipitated with anti-IKKγ antibodies. Subsequently, phosphorylation of GST-I-κBα at ser residues 32 and 36 was assessed as a marker of IKK activation. The incubation of IKK complexes with GST-IκBα in the presence of ATP exhibited minimal IKK activation in control cells or cells exposed to CTS alone. Within 10 min of IL-1β treatment, IKK induced robust phosphorylation of GST-I-κBα which was sustained until 90 min. However, in cells concomitantly exposed to CTS and IL-1β, a marked suppression of IL-1β-induced IKK activation and thus inhibition of phosphorylation of GST-IκBα was observed at all time points tested (Fig. 9 B). These observations indicated that CTS represses IKK activity.

CTS inhibits phosphorylation, degradation, and synthesis of I-κBα

Since IKKβ phosphorylates ΙκΒα for its ubiquitination and eventual proteosomal degradation, we next confirmed whether CTS mediated inhibition of IKK activation leads to inhibition of ΙκΒα phosphorylation and degradation. Cells subjected to IL-1β alone or

CTS and IL-1β for 10, 30, 60 or 90 min, were lysed and the levels of phosphorylated

ΙκΒα at serine residues 32 and 36, examined by western blot analysis. IL-1β induced phosphorylation of ΙκΒα at all the time points, flagging it for degradation. CTS however, inhibited the IL-1β-dependent phosphorylation of ΙκΒα by 62%, 88%, 95%, and 95%, at

10, 30, 60, and 90 min respectively (Fig. 9 B). Subsequently, Western blot analysis revealed that the IL-1β-induced increase in phosphorylation of ΙκΒα was followed by a rapid degradation of ΙκΒα during the first 30 min (Fig 9 B). However, inspite of an increase in the phosphorylation of ΙκΒα at 60 and 90 minutes, an increase in total ΙκΒα levels was observed due to its active synthesis (Fig 9 B, D) and continued degradation

(Fig 9 B, D). Contrarily, simultaneous exposure of cells to IL-1β and CTS, resulted in a marked suppression of IL-1β induced IκBα degradation during the initial 30 min, however, IkBα levels decreased in the cytoplasm during the ensuing 60 min (Fig. 9 B-D).

ΙκΒα has a remarkable functional diversity in the regulation of NF-κB transcriptional activation. In addition to sequestering the NF-κB in the cytoplasm, ΙκΒα shuttles

NF-κB from the nuclear to cytoplasmic compartments of the cell, to terminate its transcriptional activity. Therefore, we next determined the relative presence of ΙκΒα in the cytoplasmic and nuclear compartments of cells by immunofluorescence. FCs following

exposure to IL-1β alone demonstrated a rapid degradation of cytoplasmic ΙκΒα in IL-1β

treated cells during initial 30 min and its subsequent reexpression and degradation continued during the next 60 min (Fig. 9 C, D). Contrarily, FCs subjected to CTS and IL-1β, demonstrated inhibition of IL-1β-dependent ΙκΒα degradation in the cytoplasm by CTS during initial 30 min. Subsequently, a reduction in the total ΙκΒα was observed in cells exposed to IL-1β as well as to CTS and IL-1β (Fig 9 C, D). Analysis of immunofluorescence in FCs following simultaneous application of CTS and IL-1β, reflected a rapid and significant upregulation of intranuclear ΙκΒα within 10 min, and its translocation to perinuclear area in the next 20 min, followed by its decrease in the nucleus and cytoplasm at later time points (Fig. 9 C, D). Confirmation by Western blot analysis of the cytoplasmic and nuclear proteins further supported these findings (Fig 9 D). These observations indicated that CTS upregulates ΙκΒα nuclear import.

Since IκBα has the κB response elements in its promoter region, and its transcription is regulated by NF-κB, we speculated that the loss of IκBα at 60 and 90 min in above experiments was due to inhibition of its synthesis by CTS (118, 119). Examination of ΙκΒα mRNA expression in response to IL-1β demonstrated a rapid increase of ΙκΒα mRNA, as compared to untreated control FCs. More importantly, CTS suppressed IL-1β- mediated increase in ΙκΒα mRNA expression by 78%, 71% and 70% at 30, 60, and 90 min, respectively (Fig. 9 E). These observations collectively indicated that CTS suppresses IL-1β dependent ΙκΒα phosphorylation, degradation, and synthesis.

CTS inhibits IL-1β induced degradation of IκBβ

IKKβ activation by IL-1β also leads to IκBβ phosphorylation for its ubiquitination and degradation (118-120). IκΒβ, by sequestering NF-κB in the cytoplasm plays a small but significant role in the NF-κB canonical signaling cascade. Therefore, we investigated the consequences of the inhibition of IKK activation by CTS on the IκBβ proteolytic degradation. Western blot analysis of FCs subjected to IL-1β, revealed a slow but sustained degradation of the ΙκΒβ during the entire 10 to 90 mins (Fig 10 A). On the contrary, cells concomitantly exposed to IL-1β and CTS, demonstrated ΙκΒβ to be at levels similar to untreated control FCs, suggesting that CTS abrogated IL-1β-induced IκBβ degradation at all the time points tested (Fig. 10 A). Similarly, ΙκΒβ levels remained unchanged in FCs exposed to CTS alone. Immunofluorescence analysis further confirmed that CTS repressed IL-1β-induced intracellular degradation of IκBβ (Fig. 10 B). Although IκBβ promoter regions exhibit a κB consensus sequence, NF-κB fails to transcriptionally activate the IκBβ gene transcription (121). However, following degradation of cytosolic

IκBβ in response to IL-1β, an increase in its transcriptional activity is observed (119,

122).

CTS represses NF-κB p65 transactivation and nuclear translocation

To gain insight into the mechanisms of CTS mediated suppression of NF-κB signaling,

we further examined effects of CTS on IL-1β-induced NF-κB transactivation. Upon acti-

vation by IL-1β, IκB kinases phosphorylate Serine536, whereas phosphorylation of p65 at Serine276 by PKA is induced by degradation of the IκB proteins (119, 123). This phosphorylation of NF-κB p65 is a prerequisite for its transactivation in the nucleus.

Western blot analysis of FCs showed that IL-1β induced rapid and sustained phosphory-

lation of NF-κB p65 at Serine536 and Serine276 between 10 and 90 min (Fig. 11 A). In

FCs, IL-1β-induced phophorylation of NF-κB p65 at Serine536 was inhibited by CTS at

30 and 90 min (Fig 11 A), whereas phosphorylation of Ser276 was not regulated by CTS.

We next determined whether CTS represses IL-1β-induced NF-kB p65 binding to the DNA. In these experiments, FCs subjected to IL-1β, or IL-1β and CTS for 10 to 90

min were lysed and the nuclear proteins analyzed by EMSA. As shown in Fig 11 B, nu-

clear extracts from cells treated with IL-1β alone demonstrated progressively increased

binding of NF-kB p65 to DNA, between 10 to 90 min. CTS markedly inhibited IL-1β-

induced NF-κB p65 binding to its consensus sequences during the entire period exam-

ined. Further examination of the subunit structure of NF-κB regulated by CTS revealed

that CTS abrogated binding of NF-kB heterodimers composed of p65 (Fig 11 B). How-

ever, DNA binding of other subunits of NF-kB, such as Rel B, c-rel, p105, or p52 was not

found to bind DNA in response to IL-1β and/or CTS (data not shown).

CTS regulates IκBα nuclear transport

Unphosphorylated form of IκBα enters the nucleus to bind NF-κB and its nuclear export signal allows transport of nuclear NF-κB-IκBα complexes back to the cytosol, thereby

terminating its transcriptional activity (118, 119). In above studies we observed that fol-

lowing simultaneous exposure to CTS and IL-1β, (i) IκBα rapidly transmigrates to the

nucleus within 10 min, and (ii) NF-κB that translocates to the nucleus during initial 10 to

30 min, does not bind to its consensus sequences. In order to visualize the temporal regu-

lation of ΙκΒα and NF-κB in the cytoplasmic and nuclear compartments, FCs were acti-

vated as described above and the presence of NF-κB p65 and ΙκΒα in the nuclear and

cytoplasmic compartments examined in the same cells by immunofluorescence. Follow-

ing exposure to IL-1β alone, a rapid degradation of cytoplasmic ΙκΒα was paralleled by a

rapid translocation of NF-κB to the nucleus (Fig. 11 D). FCs subjected to CTS and IL-1β

demonstrated that CTS inhibited IL-1β-dependent ΙκΒα degradation, however, it induced

a rapid upregulation of intranuclear ΙκΒα during the initial 10 min. Simultaneous visuali-

zation of NF-κB in these cells revealed retention of cytoplasmic NF-κB, even though

translocation of NF-κB to a lesser extent in the nuclei was apparent during the initial 10

min (Fig 11 D). As observed above, examination of FCs after ensuing 20 min of IL-1β

treatment, exhibited negligible presence of ΙκΒα in the nuclei or cytoplasm, and in parallel nuclear translocation of nearly all NF-κB. Contrarily in FCs treated with CTS and IL-

1β, majority of the ΙκΒα and NF-κB were present in the perinuclear area and cytoplasm, with presence of remaining ΙκΒα and NF-κB in the nuclei. Interestingly, following 60 and 90 min of activation of FCs with IL-1β, the levels of ΙκΒα and NF-κB progressively increased in the nucleus and in the cytoplasm. Noticeably, in FCs exposed to IL-1β and

CTS simultaneously, the ΙκΒα levels were lower in the cytoplasm, with a clear lack of intranuclear ΙκΒα and NF-κB by 90 min. These observations further demonstrated that

CTS upregulates ΙκΒα nuclear trafficking to facilitate NF-κB export from the nucleus to terminate NF-κB transcriptional activity.

To further confirm the relevance of ΙκΒα nuclear translocation in NF-κB signaling, the nuclear extracts from FCs subjected to IL-1β in the presence or absence of CTS were subjected to co-immunoprecipitation with anti IκBα and anti NF-κB p65 immu- noglobulins, and analyzed by Western blots. The presence of NF-κB or IκBα was negligible in untreated control cells as well as cells exposed to CTS alone. FCs treated with

IL-1β exhibited presence of nuclear NF-κB and IκBα but insignificant NF-κB-IκBα complexing during the first 30 min of activation, and their subsequent upregulation in nuclear compartment at 60 and 90 min. More importantly, simultaneous exposure of CTS and IL-1β demonstrated a dramatic upregulation of NF-κB-IκBα complexes in nuclei during first 30 min and their subsequent decrease in the ensuing 60 minutes (Fig. 11 C).

Collectively, these findings strongly demonstrate that CTS, in addition to inhibiting

TAK1 activity, induces a rapid IκBα nuclear translocation to facilitate quick export of

nuclear NF-κB to prevent its transcriptional activity and subsequent proinflammatory

gene induction.

Discussion

Our observations demonstrate that tensile forces of low/physiologic levels attenuate inflammatory gene induction by three major proinflammatory molecules, IL-1β, TNF-α, and LPS, individually and collectively, by inhibiting a critical event in the NF-kB activation. All of these mediators act on cells via discrete receptors to initiate inflammatory responses (119, 122, 124). Signals generated by CTS do not downregulate receptors for IL-

1β, TNF-α, or LPS as a mechanism to inhibit proinflammatory signaling cascade.

Signals generated by major inflammatory ligand specific receptors converge at

TAK1 via activation of discrete intermediate mediators and co-adaptors. CTS inhibits IL-

1β, TNF-α, or LPS induced phosphorylation of TAK1 at Thr187 necessary for its kinase activity. The reduction in TAK1 activation substantially decreased IL-1β-dependent induction of IKK activity and phosphorylation of ΙκΒα at serine residues 32 and 36 by the

IKK complexes. This proved that CTS regulates TAK1 as a critical control point to limit proinflammatory gene induction (Fig 5).

Inhibition of IKK activity prevents IκBα and IκBβ phosphorylation and degradation, thus prolonging IκBα and IκBβ mediated cytoplasmic sequestration of NF-κB. CTS also downregulates IκBα gene transcription, further repressing activation of NF-κB.

Nevertheless, CTS was found not to completely block the synthesis of IκBα or IκBβ beyond levels of constitutive synthesis. Thus a pool of IκBα and IκBβ was always present in the cells, that contributed to sequestration of the NF-κB in the cytoplasm of cells after

60 min of CTS exposure.

Another way in which NF-κB activity could be potentially inhibited is by shuttling of NF-κB from the nucleus to terminate its transcriptional activity (119, 122, 124).

In addition to inhibiting nuclear translocation of P65, CTS also abrogates DNA binding of P65 that rapidly translocates to the nucleus at the earlier time points. A nuclear export signal on ΙκΒα allows shuttling of nuclear NF-κB to the cytoplasm, thereby providing yet another point of control over NF-κB transcriptional activity (118, 119). CTS regulates shuttling of IκBα to further terminate NF-κB activity. During first 30 min IL-1β activation, minimal nuclear import of ΙκΒα is observed. Strikingly, CTS upregulates rapid nuclear import of ΙκΒα and its binding to NF-κB p65 within the first ten minutes, to export any NF-kB dimers that enter the nucleus. By 60 min the nuclear levels of IκBα are decreased by CTS, which could be attributed to the reduction in nuclear translocation of

NF-κB, resulting in a reduction of IL-1β induced ΙκΒα synthesis (Fig 5). Thus, CTS suppresses the IL-1-induced transcriptional activation of NF-κB at multiple levels by in-

hibiting TAK1 activation, suppressing IκBα degradation, inhibition of IκB gene transcription, and promoting its nuclear shuttling to terminate NF-kB DNA binding.

To generate inflammation cells have the ability to respond to different ligands in distinct ways. Mechanical signals apparently employ distinct mechanisms to regulate cellular responses in order to terminate NF-κB activity, which are critical for generating functional antiinflammatory effects. Inflammatory disorders of the joints are convention- ally treated with various inhibitors of proinflammatory mediators. Recent efforts have sought to block critical steps in NF-κB signaling cascade to alleviate arthritic pain and cartilage destruction. However, all the measures have shown limited success and are ac- companied by undesirable side effects. CTS is an effective non-pharmacological signal that acts as a molecular switch of NF-κB activity to inhibit inflammation in fibrochondrocytes. This report provides the mechanisms underlying the well recognized but little understood role of mechanical signals in ameliorating inflammation of the joints.

Acknowledgements. This work was supported by the National Institutes of Health by grants AR04878, DE15399, AT00646, and HD40939.

FIGURE 8. CTS inhibits IL-1β, TNF-α, and LPS induced proinflammatory mRNA expression. (A) NOS2A or (B) MMP-13 mRNA expression in FCs treated with IL-1β,

TNF-α, and LPS with or without simultaneous exposure to CTS for 3 h. (C) NOS2A protein expression in control FCs, in FC exposed to CTS or stimulants (IL-1β, TNF-α or

LPS) alone, or with stimulants after one h prexposure to CTS, or simultaneously to stimulant and CTS. The induction of NOS2A shows that preexposure of FCs to CTS does not inhibit its responsiveness to the stimulants. * indicates p< 0.5

Figure 8.

FIGURE 9. CTS suppresses IL-1β-induced TAK1 activation and IκBα activation. (A)

Phosphorylation of TAK1 Threonine187 in IL-1β, TNF-α or LPS-activated FCs and its

inhibition by CTS during the initial 10 min of activation. (B) Phosphorylation of GST-

IκBα by IKK complexes immunoprecipitated with anti-IKKγ antibodies from lysates of

cell exposed to IL-1β in the presence or absence of CTS for 10 to 90 min. Western blots were probed with anti-phospho-I-κBα mouse monoclonal antibody (p-I-κBα), and anti-

N-terminal-I-κB-GST IgG as I-κB loading control. The total IKK levels were assessed by

probing the blots with rabbit anti-IKKγ IgG. The densitometric imaging with infra-red

IRdye 680 or 800 labeled secondary antibodies shows that CTS inhibited IL-1β-induced

activation of IKK. In same cells, Western blot analysis showed that CTS inhibited IL-1β-

induced phosphorylation of I-κBα at Serine32 and -36, degradation and synthesis. (C)

Immunofluoresence analysis with anti-IκBα antibodies showing that CTS inhibits IL-1β-

dependent degradation and synthesis of I-κBα. The nucleus has been stained with DAPI

and I-κBα has been shown in red using CY3. Pink areas in the nucleus indicate increased

presence of I-κBα. (D) Western blot analysis of cytoplasmic and nuclear I-κBα levels in

FCs of I-κBα showing a CTS-induced influx of I-κBα into the nucleus within 10 min and

inhibition of its degradation in the cytoplasm. (E) I-κBα mRNA expression in FCs exhib-

iting inhibition of IL-1-dependent I-κBα mRNA expression by CTS. Data represent one

of three independent experiments with similar results.

Figure 9

Figure 9 continued

FIGURE 10. CTS inhibits IL-1β-induced I-κBβ degradation. (A) Western Blot analysis showing the inhibition of IL-1-induced I-κBβ degradation by CTS. (B) Immunofluores- cence analysis demonstrating inhibition of I-κBβ degradation by CTS.

Figure 10

FIGURE 11. CTS inhibits IL-1β-induced NF-κB phosphorylation, nuclear translocation and DNA binding. (A) Western blot analysis of FC lysates showing CTS inhibits IL-1β- induced phosphorylation of NF-κB p65 at Serine536 at 30 and 90 minutes as assessed by densitometric imaging of IRdye 680 labeled secondary antibodies. (B) EMSA analysis exhibiting DNA binding of NF-κB p65 from the nuclear extracts of FCs showing lack of

DNA binding in untreated control FCs (Lane 1); increased binding of NF-κB p65 in IL-

1β treated cells (Lanes 2, 4, 6, and 8); and inhibition of NF-κB binding to DNA in cells treated with CTS and IL-1β (lanes 3, 5, 7, and 9). IL-1β treated cells incubated with 100- fold excess unlabeled probe (lane 10), cells exposed to IL-1β and control IgG (Lane 11),

CTS alone (lane 12), and super shift EMSA showing CTS regulates NF-κB subunit p65 to DNA (Lane 13). (C) Complexing of I-κBα with NF-κB p65 following treatment of

FCs with IL-1β in the presence or absence of CTS was analysed in the immunoprecipi- tates of IκBα. Western blot analysis showed the binding of NF-kB to IκBα in cells exposed to CTS and IL-1β and a lack of IκBα-NF-κB complexes during the initial 10 min of activation. Additionally, presence of IκBα-NF-κB complexes was apparent in IL-1β- treated cells at 60 min, and a lack of such complexes in FCs exposed to CTS and IL-1β.

(D) Immunostaining of NF-κB (green) and IκBα (red) and their merged images in FCs treated with IL-1β (Left panel) or CTS and IL-1β (right panel) at 10, 30, 60 or 90 min.

FCs exhibit that IL-1β induces rapid and sustained translocation of NF-κB to the nucleus.

CTS induces rapid translocation of IκBα to the nucleus, and increases complexing with the minimal NF-κB that enters into the nucleus during first 10 min due to the rapid actions of IL-1β. This IκBα shuttles the nuclear NF-κB out, as evidenced by the presence of yellow stain in the nuclei and the presence of IκBα-NF-κB complexes in the cytoplasm. The IκBα-NF-κB complexes are also evident after 30 and 60 min in the cytoplasm of the cells treated with CTS and IL-1β, suggesting nuclear export of NF-κB by

IκBα. This confirms the findings from (B) and (C). The data is representative of 3 separate experiments.

Figure 11

Figure 11 continued

100

FIGURE 12. Schematic representation of the mechanisms of intracellular actions of

CTS showing CTS suppresses IL-1β induced proinflammatory gene induction by inter- cepting salient steps in the NF-κB signaling cascade to inhibit its transcriptional activity by: (i) suppressing IL-1β-induced TAK1 activation and thus phosphorylation and activation of IKK. The suppression of IKK activity results in the inhibition of phosphorylation and proteosomal degradation of I-κBα and I-kBβ, thus inhibiting nuclear translocation of

NF-κB. (ii) During the initial 30 minutes, CTS upregulates I-κBα nuclear translocation to prevent NF-κB binding to the DNA and facilitates export of nuclear NF-κB, that may enter the nucleus due to rapid actions of IL-1β. (iii) CTS represses IL-1β-induced I-κBα and I-κBβ mRNA expression to the control levels and also prevents NF-κB DNA binding. Collectively, these actions of CTS inhibit proinflammatory gene induction and molecules involved in the regulation of NF-κB signaling cascade. The red dots indicate steps within the NF-kB pathway that are inhibited by CTS. The green dots indicate molecular events within the NF-kB pathway that are augmented by CTS.

101

Mechanical signals IL-1β

LPS

TNF-α TAK1

IKK (NF-κB)/(IκBβ) p-IκBβ P (IκBα)/(NF-κB)c (IκBα)c P p-IκBα

(IκBα)n (IκBα)n I-mRNA

p-NF-κB P-mRNA

Figure 5

Figure 12

102

CHAPTER 5

DIFFERENTIAL REGULATION OF PROINFLAMMATORY

GENES BY MECHANICAL SIGNALS LEADS TO

ALVEOLAR BONE REMODELLING

INTRODUCTION

Biomechanical signals play a vital role in bone resorption and bone deposition. However

the intracellular mechanisms by which mechanical signals modulate bone remodeling

remain mostly obscure. Bone is a mechano-sensitive tissue and adapt to various types and

magnitudes of mechanical forces. It has been well recognized that compression leads to

bone resorption, whereas bone deposition occurs at sites of optimal magnitudes of tension

(126, 127). Such responses of bone to mechanical forces are frequently observed during bone adaptation to distraction, fracture healing, tooth movement, and application of load

(128, 129). Both osteoblasts and osteocytes are shown to respond to mechanical signals.

For example, mechanical loading increases expressions of collagen type I, dentin matrix

103

protein (DMP-1), matrix extra-cellular phospho-glycoprotein (MEPE), and insulin-like growth factor (IGF-1), in osteocytes. Similarly, these signals upregulate the expression of osteocalcin and alkaline phosphatase in osteoblasts and osteoblast-like periodontal ligament (PDL) cells (130-132). The effects of mechanical forces on cells are mediated by multiple secondary messengers such as, cyclic AMP, inositol phosphates, nitric oxide, G proteins, tyrosine kinases, and calcium that are increased or decreased in response to mechanical signals (133-135). Additionally, inflammatory mediators like interleukin-1 (IL-

1) and tumor necrosis factor-α (TNF-α) play a significant role in bone adaptation. Stimu- lation of the arachidonic acid metabolism and subsequent production of prostaglandins have also been attributed as a mechanism for bone formation as well as resorption (136,

137).

Bone homeostasis is a dynamic process in which mechanical signals are vital for both bone formation and resorption. This phenomenon is clearly depicted during tooth movement where mechanical signals activate alveolar bone remodeling. The PDL cells juxtaposed between teeth and bone play a critical role in bone remodeling. In response to compressive forces, these cells produce mediators such as prostaglandins (PGE2), interleukin-1 and tumor necrosis factor-a (TNF-α), which diffuse all around the roots in PDL

(128, 139). Consequently, bone resorption is observed at sites experiencing compressive and high tensile forces, whereas bone deposits in fields experiencing tensile forces of low magnitudes. Yet, how the bone forms in tensile fields despite the presence of proinflammatory mediators is quite intriguing. Earlier we have shown that tensile forces of low 104

magnitudes (LTF) are anti-inflammatory and induce anabolic signals in osteoblast-like

periodontal ligament cells, in a magnitude dependent manner and regulate inflammatory

gene transcription, in-vitro (140, 141). In contrast, tensile forces of high magnitudes

(HTF) act as pro-inflammatory stimuli and increase the expression of inflammatory cytokines.

The primary objective of this study was to present a mechanistic explanation for the bone resorption and bone deposition in response to mechanical signals by correlating the in-vitro anti-inflammatory actions of low tensile forces with the suppression of inflammation in-vivo. To this end, we utilized a rat model of tooth movement to assess the mechanisms by which tensile forces of high and low magnitudes regulate expression of

COX2 and IL-1 to regulate bone formation in periodontal tissues. In this model the strain experienced by the PDL of a root subjected to orthodontic forces is maximal at the cervical region and gradually reduces along the root length, to reach values closer to zero at the root apex (142, 143). Thus, this model provides an ideal means to understand the effects of biomechanical forces on bone remodeling. In parallel, we examined the effects of tensile forces on osteoblast-like PDL cells in vitro, to investigate the intracellular mechanisms by which tensile forces inhibit inflammatory signals and induce bone formation.

105

MATERIALS AND METHODS

Tooth mobilization and Immunostaining

The upper first molars of 45 days old male Sprague-Dawley rats (n=12) were tipped mesially using two springs with a built-in force of 40 cN for 48 hours. One end of the spring was bonded to each molar through custom-made cleats and the other end was fixed to the pinned incisors, which were used as anchors. After euthanizing, the hemimaxillae were fixed, decalcified and sectioned horizontally from cervical to apical levels. The detailed methods for this tooth movement animal model and specimen preparations have been published elsewhere (144, 145). The sections of the roots of first molars from the cervical region and from region close to apex were immunohistochemically stained to examine the presence of COX-2 or IL-1β, using rabbit anti-rat COX-2 IgG (Cayman Chemicals,

MI) or rabbit anti-rat IL-1β IgG (Abcam, MA) as primary and CY3 or CY2 conjugated goat anti-rabbit IgG (Jackson Immuno Research, PA) as secondary antibodies.

Isolation of Human PDL cells

PDL cells were harvested from the surfaces of the roots of healthy, erupted human third molars that were extracted for orthodontic reasons. The cells were then cultured in tissue culture medium (TCM) containing Basal Eagle's Medium (Invitrogen, CA), 10% low en- dotoxin FCS (Atlanata Biologicals, GA), 2 mM glutamine, 100 U/ml penicillin, 80 µg/ml tylosin and 100 µg/ml streptomycin (Invitrogen, CA). The semi-confluent cultures were

106

grown further in tylosin-free medium, cloned by limiting dilutions, and were used for experiments between 6th and 20th passages, as described earlier (140, 141, 146). PDL cell clones, designated as PL-150 and PL-75 (from white females, 18 yrs) and PL-484 (from a white male, 22 yrs), retained their osteoblast-like phenotype as shown by the expression of collagen type-I, TGF-β, alkaline phosphatase, and Runx-2 (140, 141, 146).

Application of equibiaxial strain (TF) on PDL cells in vitro

PDL cells were seeded on Bioflex II culture plates (5 x 105/well), grown to 70% confluence (6-8 days old), washed, and incubated in serum free TCM overnight. Subsequently, culture plates were subjected to equibiaxial strain in a Flexercell 4000 Unit (Flexcell In- ternational Corp, NC) at a magnitude of 6% and a frequency of 0.005 Hz for required periods of time (140, 141, 146). To provide uniform radial and circumferential strain on the membrane, the plates were placed on a loading station (located in a 5% CO2 incubator with 95% humidity), such that following application of vacuum, the membrane deformed across the post-face creating uniform equibiaxial strain. The strain was calculated as:

π × ()change2 radiusin (change radiusin ) circumferential strain = = = radial strain 2π × ()original radius ()original radius

Real time reverse transcriptase/polymerase chain reaction

RNA was isolated from cells using the RNeasy kit (Qiagen, CA) after shredding through a Qiashredder (Qiagen, CA). The RNA was subjected to DNAse digestion, and stored in

107

40 μl of RNAse-free water. The concentration and purity of RNA was spectrophotometri- cally measured at 260 nm. A total of 1 μg of RNA was mixed with 1μg oligo-dT (12-18 oligomers) in RT buffer and incubated for 10 min at room temperature. This mixture was transcribed with 200 units of MULV reverse transcriptase for 30 minutes at 37°C. The cDNA was amplified with 0.1 μg of specific primers in a reaction mixture containing 200

μM dNTP and 0.1 units of Taq polymerase in PCR buffer. Gene specific primer sequences were selected using the Taqman Probe and Primer Design function of the Primer

Express v1.5 software (Applied Biosystems, CA). The sense and antisense sequences of human primers used were as follows: IL-1β sense 5’-

CTGAGCACCTTCTTTCCCTTCA-3’, antisense 5’-

ATCGTGCACATAAGCCTCGTT3’, probe 6-

FAMd(AACCTATCTTCTTCGACACATG)BHQ-1; COX-2 sense 5’-

CTTTTGGTGGAGAAGTGGGTTT-3’, antisense 5’-

GACAGCCCTTCACGTTATTGC-3’, probe 6-FAMd(TCAACACTGCCTCAATT)BHQ-1; GAPDH sense 5’-

AATTCGCACGAGGCTTCTTTT -3’, antisense 5’-CCGTTGACTCCGACCTTCAC-3’, probe 6-FAMd(AGCCACATCGCTCAGA)BHQ-1. Reverse transcription reactions were carried out using 2 µg RNA and TaqMan Reverse Transcription reagents, followed by real-time PCR using Taqman® PCR Master Mix and ABI Prism 7700 Sequence Detec- tion System® (Applied Biosystems, CA). Reactions were performed as follows: Cycle I

(1X): 95º C for 3.0 min, Cycle II (50X): Step 1 at 95º C for 0.3 min, followed by Step 2

108

at 55º C for 0.3 min, and Step 3 at 72º C for 30 mins, Cycle III at 40º C hold. Following

amplification, a melting curve was obtained to ensure that primer-dimers or non-specific

products had been eliminated or minimized. The data, obtained by real-time PCR, was

analyzed by the comparative threshold cycle (CT) method. In this method, the amount of

the target, normalized to GAPDH, and relative to a calibrator (either untreated sample or

IL-1β-stimulated cells), is given by 2-∆∆CT, where ∆∆CT = ∆CT (sample) - ∆CT (calibrator), and ∆CT is the CT of the target gene subtracted from the CT of GAPDH.

In some cases the expression of mRNA was assessed by end point PCR (RT/PCR)

as described earlier (140, 141, 146). The sense and antisense sequence of human primers

used were as follows: GAPDH (548 bp): sense 5’-GGTGAAGGTCGGAGTCAACGG-

3’, and antisense 5’-GGTCATGAGTCTTCCACGAT-3’, Runx-2: Sense: 5’-

AGCCTCTTCAGCGCAGTGAC-3’ antisense 5’-CTGGTGCTCGGATCCCAA-3’; al-

kaline-phosphatase: Sense: 5’-CGTGGCCAAGAACATCATCA-3’, antisense: 5’-

GCGGGCAGCTGTCACTGT-3’, Osteopontin: Sense: 5’-

TGAGACTGGCAGTGGTTTGC-3’, antisense: 5’-CCACTTTCACCGGGAGACA-3’.

PCR products were separated on a 2%-agarose gel at 100V/cm in Tris-acetate/EDTA

electrophoresis buffer. The gels were stained with ethidium bromide and then photo-

graphed in Kodak Image Station 440 CF. The intensity of the bands was analyzed by

IMAGE J (NIH, MD) program.

109

Immunofluoresence

Following exposure of cells to various treatment regimens, the membranes were fixed with 2% paraformaldehyde. The expression of proteins was analyzed by immunostaining, using rabbit anti-NF-κB p65 IgG (Santa Cruz SC-109), rabbit anti-I-κBβ IgG (Santa

Cruz Biotech, CA) and using secondary antibodies, goat anti-rabbit-CY3 and donkey anti-goat-FITC (Jackson Lab, MN). Subsequently, the membranes were mounted on slides with Vectashield (Vector Labs, CA), and observed under an epifluoresence microscope

(Ziess Axioimage) or by laser scanning cytometry (LSC). At least 3 membranes from separate experiments were analyzed. On each membrane, 8 to 10 areas of 100 cells each were counted to assess the number of fluorescence positive cells by LSC or intensity of fluorescence by Zeiss Axiovision software (Carl Zeiss, Germany).

Western blot analysis

For semiquantitative measurements of protein synthesis Western blot analysis was used as described earlier (137-139). Briefly, cells were lysed in ice cold RIPA buffer (Santa

Cruz Biotech, CA) containing protease inhibitor cocktail (Roche, IN), and the extracted proteins were loaded on the SDS-10% acrylamide gels. The proteins were electrophoretically transferred to nitrocellulose membranes (Bio-Rad, CA) and identified by rabbit anti-

I-κBβ IgG (1:600 dilution), and rabbit anti-NF-κB p65 IgG (Santa Cruz Biotechnology,

CA). Monoclonal mouse anti-β-actin IgG (1:20,000; Abcam, MA) was used to re-probe the same blots to equilibrate protein input in all lanes. Horseradish peroxidase-labeled 110

(HRP) donkey anti-mouse IgG (1:10,000 dilution; Chemicon, CA) or HRP-labeled goat anti-rabbit antibody (1:10,000 dilution; Santa Cruz Biotechnology, CA) was used as a second antibody. The presence of HRP was detected by Luminol (Amersham, IL), and the semiquantitative analysis of luminescent bands was carried out with Kodak Image

Station 1000®, and Kodak 1D image analysis software.

Data analysis and statistics

The data represents means and standard deviations of three experiments performed in triplicates. At least two sections of hemimaxillae were used for each analysis. The SPSS

13.0 software (SPSS Inc., Chicago, IL) was used for statistical analysis. To identify differences between IL-1β treated cells in the presence of LTF, One-Way ANOVA and the post-hoc multiple comparison Dunnett test was used. p<0.05 was considered statistically significant.

RESULTS

Differential expression of IL-1β and COX-2 at sites exposed to compressive or tensile forces

In order to examine the effects of varying magnitudes of mechanical forces on the PDL

(Fig. 13 D), the first molar roots of rats were subjected to orthodontic forces of 40 cN and 111

the expression of COX-2 and IL-1β analyzed by immunostaining of the sections (144,

145). As illustrated in Fig. 13 D, the tipping action of the orthodontic forces resulted in

the maximal force generation at the cervical region at both compressive and tensile fields,

which gradually reduced along the root length, with the least forces experienced at the

apex. The sections closer to the cervical region of the first molar roots exhibited abundant

expression of COX-2 and IL-1β at both tension (T) and compression (C) sites that were

exposed to excessive applied forces (Fig 13 A and B). Interestingly, the cells of the pulp

showed abundant expression of IL-1β and COX-2, indicating the cells of the pulp also

experienced high levels of forces. The sections closer to the apical region, that were ex-

posed to lower levels of forces showed minimal expression of COX-2 and IL-1β at ten-

sion sites, but high expression of COX-2 and IL-1β at compression sites. Negative con-

trol sections taken through cervical region of the third molar of the same maxilla, did not

exhibit expression of COX-2 and IL-1. The insets of Fig 1A and B show higher magnifi-

cation of COX-2 and IL-1β staining at tension (T) and compression (C) sites. These ob-

servations support the fact that compressive and tensile forces of higher magnitudes act

as a pro-inflammatory signal, whereas tensile forces of lower magnitudes either do not

induce expression and/or may suppress the expression of pro-inflammatory molecules

present in the PDL. We further analyzed the entire hemimaxilla from the 1st to 3rd molar regions to ascertain that the expression of proinflammatory mediators was dependent on the orthodontic forces applied. As shown in Fig. 13 C, pro-inflammatory mediators exhibited a spatial expression from the first molar to the third molar region. As evident, the 112

forces applied to the first molar were also experienced by the cervical regions of the sec-

ond molars. However, the third molars did not exhibit expression of IL-1β or COX-2 at

compressive or tension sites.

Tensile forces regulate proinflammatory gene transcription in PDL cells in vitro

COX-2 and IL-1β are the most common mediators of inflammation in the PDL. These

mediators were increased at the sites of bone resorption in the PDL (Fig. 13 A) and have

been reported to be present all around the roots due to their diffusion in the PDL (137,

138). Therefore, to further determine the effects of tensile forces of low magnitudes

(LTF) on IL-1β-induced proinflammatory gene expression, cells were subjected to four

different treatment regimens, control untreated cells, cells exposed to IL-1β (1 ng/ml),

cells subjected to LTF alone (6% TF at 0.005 Hz), or cells subjected to LTF and IL-1β,

for 24 h. As evident in Fig. 14 A, cells exposed to IL-1β showed a marked increase in

COX-2 and IL-1β mRNA expression as compared to control cells over a period of 24 h.

However, LTF blocked IL-1β-induced COX-2 (92%) and IL-1β (86%) mRNA expression significantly (p ≤ 0.001). In these experiments LTF alone did not induce COX-2 or IL-1β

expression.

We next examined the effects of tensile forces of higher magnitudes (HTF) on the expression of IL-1β and COX-2 in PDL cells. As shown in Fig. 14 B, HTF (15% TF at

0.005Hz) induced a significant upregulation of IL-1β expression in PDL cells, in the ab-

sence of IL-1β. For example, the induction of COX-2 mRNA was 1122 % greater in 113

cells exposed to IL-1β and HTF as compared to untreated control cells, and 180% greater in cells exposed to HTF alone as compared to control cells. Similarly, HTF alone induced a 540% higher expression of COX-2 as compared to control cells, and HTF + IL-1β in-

duced a 1060% greater expression of IL-1β in comparison to untreated control cell.

LTF exerts anti-inflammatory actions by modulating the NF-kB pathway

Since the pro-inflammatory mediators like COX-2 and IL-1β are regulated by the NF-κB

transcription factors, we next examined whether tensile forces regulated the mRNA tran-

scription of these genes via NF-κB pathway (147). In these experiments PDL cells were

treated with the four different regimens for 30 min and the nuclear translocation of NF-

κB examined by immunofluorescence. As shown in Fig. 15 A, the control untreated cells

exhibited NF-κB predominantly in the cytoplasm, whereas exposure of cells to IL-

1β resulted in a rapid nuclear translocation of NF-κB. However, in cells treated with IL-

1β and LTF simultaneously, NF-κB was retained in the cytoplasm similar to that ob-

served in control cells.

Examination of the mRNA expression for NF-κB subunit p65 revealed that IL-1β

upregulated the expression of p65 within 15 minutes and its induction was sustained for

the ensuing 75 minutes, as compared to control cells. However, LTF suppressed IL-1β-

induced NF-kB mRNA expression between 30 and 120 mins (Fig. 15 B). We further ex-

amined whether the LTF-induced inhibition of NF-κB mRNA was reflected in NF-κB

114

p65 synthesis. As shown in Fig. 3C, cells subjected to IL-1β exhibited a marked upregu-

lation of NF-κB synthesis. On the contrary, LTF repressed more than 90% of the IL-1β-

dependent NF-κB p65 synthesis in a sustained manner between 30 and 120 minutes. In these experiments, LTF alone did not regulate NF-κB mRNA expression or its synthesis

(Fig. 15 C).

LTF inhibits nuclear translocation of NF-kB by inhibiting I-κB degradation

I-κB proteins play a critical role in sequestering NF-kB in the cytoplasm to inhibit its nuclear translocation (147). Therefore, we next examined whether LTF inhibits I-κB degradation and /or synthesis to exert its effects on NF-kB activity. In these experiments, following treatment of PDL cells with IL-1β or IL-1β and LTF for various time intervals,

the presence of I-κBα and I-κBβ was analyzed by Western blots. As evident in Fig. 16 A

and B, IL-1β caused degradation of I-κBβ within 30 minutes that continued for the ensu-

ing 90 minutes. This IL-1β induced degradation of I-κBβ was markedly inhibited by LTF

and the levels of I-κBβ were restored to steady state levels, close to those found in con-

trol cells. On the other hand, IL-1β-treated cells showed a reduction in the total I-

κBα. Interestingly, the cells treated with IL-1β and LTF showed an inhibition of I-κBα

degradation during the first 30 min, however from 60 to 120 min a reduction in total I-

κBα was observed (Fig. 16 A and B). In these experiments, LTF alone did not affect in-

tracellular I-κBα pools. Since I-κBα gene expression is controlled by NF-κB transcrip-

115

tion factors, we next examined whether LTF-induced decrease in total I-κBα was due to the inhibition of NF-κB nuclear translocation that subsequently results in the inhibition of

I-κB mRNA expression and its synthesis. As expected, IL-1β treated cells showed a sub- stantial induction in the expression of I-κBα within 15 minutes. This induction in mRNA expression was maintained up to 120 minutes. LTF, however suppressed the IL-1β- mediated I-κBα mRNA expression at all time points tested, as compared to cells treated with IL-1β alone (Fig 16 C). These observations revealed that the loss of I-κBα in the

LTF and IL-1β treated cells was due to the suppression of I-κBα synthesis. Furthermore, the inability of NF-κB to translocate to the nucleus was due to the LTF-dependent inhibition of I-κBβ cytoplasmic degradation, resulting in the failure of dissociation between

NF-κB and I-κB cytoplasmic complexes.

LTF counteracts IL-1β induced suppression of genes involved in bone formation

In order to examine whether LTF, while inhibiting inflammation, also exerts anabolic gene induction in PDL cells, we next investigated the effects of LTF on the regulation of osteogenic markers like Runx-2, Osteopontin and alkaline phosphatase. IL-1β inhibited mRNA expression of osteopontin, alkaline phosphatase and Runx-2 by 75%, 78% and

90%, respectively as compared to control untreated cells. On the contrary LTF abrogated the IL-1β induced inhibition of Runx-2, Osteopontin and alkaline phosphatase and their levels were close to untreated control cells (Fig. 17 A, B and C).

116

DISCUSSION

In this study the responses of bone to biomechanical signals were investigated. Previ- ously we have established an in-vitro model in which osteoblast-like PDL cells could be exposed to varying magnitudes of cyclic tensile strain (140, 141). Utilizing this system, we have observed that cyclic tensile strains of low magnitudes are anti-inflammatory and suppress IL-1β-induced proinflammatory gene induction (140, 141). Here, in order to understand the mechanisms of bone deposition in tensile fields in vivo, we used tooth movement as an in-vivo model and in parallel investigated the intracellular mechanisms of actions of mechanical forces on human PDL cells in vitro.

It is well documented that during tooth movement bone resorbs at sites of compression and high tensile forces, and deposits at sites of appropriate tensile forces (146,

148-150). During this process, mediators such as IL-1β and TNF-α, cyclooxygenase-2

(COX-2) and its downstream eicosanoid PGE2, are shown to be present in both compressive and tensile fields in the periodontal ligament of roots (133, 136, 137). However, bone deposition and resorption are observed to be regulated by the magnitude of mechanical forces experienced by the various regions of the PDL and bone. In order to understand the mechanisms of bone deposition in tensile fields in vivo, the bilateral upper first molars of Sprague-Dawley rats were mesially tipped using an orthodontic force of

40 cN. As a result, the PDL was exposed to differential forces with the mesial surface of

117

the first molar roots being exposed to compressive forces, and the distal aspect of the root subjected to tensile forces. In this model, the rat molar root experiences high strain levels at the cervical region, which gradually decreased along the root to reach values closer to zero at the apex (142, 143). This decreasing strain gradient in the cervico-apical direction renders the tooth to varied levels of tensile forces, making it an ideal model for studying the effects of varied magnitudes of tensile forces on bone remodeling. We have observed by immunostaining of the hemimaxillae, a markedly high expression of IL-1β and COX-

2 expression in the cervical region of the tooth that corresponded to the areas experiencing compressive forces as well as tensile forces of high magnitudes. However, in these studies, the expression of both IL-1β and COX-2 was apparent only in the apical region of the roots exposed to compressive forces. More importantly, sections closer to the apex

of the roots showed a marked absence of the pro-inflammatory mediators in the areas ex-

posed to tensile forces, suggesting that either expression of IL-1β and COX-2 was re-

pressed, or these cytokines were not synthesized in the areas of roots exposed to tensile

forces of low magnitudes. The PDL is a gelatinous tissue, therefore IL-1β and COX-2

secreted at the compression sites diffuse along the roots and are also present at the sites

experiencing tensile forces (137-139). Therefore, it is reasonable to expect that the mechanical forces of low magnitude inhibited the expression of IL-1β and PGE-2 to allow

bone formation in vivo. Interestingly, the pulp of the roots at the cervical area also

showed profuse expression of both IL-1β and COX-2 . This suggests that pulp cells also

118

experience and respond to these forces and respond to them via proinflammatory gene induction.

To further understand the mechanisms of actions of tensile forces on bone formation in a proinflammatory environment we utilized the in vitro system where well- characterized PDL cells with osteoblast-like phenotype were exposed to IL-1β at concentrations that induce both IL-1β and COX-2 (140, 141). Similar to in vivo findings, these experiments also showed that signals generated by tensile forces of low magnitudes significantly inhibited IL-1β-induced expression of both COX-2 and IL-1β mRNA, while tensile forces of higher magnitudes themselves were sufficient to generate both IL-1β and

COX-2 mRNA expression ( 137, 138). Genes like IL-1β and COX-2 are regulated by the

NF-κB transcription factors and mediate inflammatory gene transcription. Examination of the nuclear translocation of NF-κB by immunostaining and Western blot analysis revealed that IL-1β induced a rapid nuclear translocation of the NF-κB subunit P65. More importantly, signals generated by LTF prevented the NF-κB p65 nuclear translocation

(140, 141, 143). In addition to an inhibition of nuclear translocation, LTF also inhibited the IL-1β induced expression and synthesis of NF-κB p65.

The NF-κB subunits are sequestered in the cytoplasm bound to the inhibitor proteins, I-kappaB (I-κB) α and β. We have observed that LTF inhibits the nuclear translocation of NF-κB P65 by inhibiting I-κB α and β degradation in order to exert its anti- inflammatory actions (38, 140, and 141). Analysis of the I-κBα degradation revealed that

119

LTF inhibits IL-1-dependent I-κBα degradation initially during the first 60 min. How- ever, during the later stages a total reduction in I-κBα in cells treated with IL-1β and LTF reflects decrease in its synthesis due to inhibition of its mRNA expression. Since NF-κB controls I-κBα mRNA transcription, it is not surprising that, while IL-1β increased the mRNA expression of I-κBα, signals generated by LTF inhibited IL-1β-induced I-κBα mRNA expression. Similar to I-κBα, I-κBβ was also degraded by IL-1β in order to facilitate NF-kB nuclear translocation. LTF inhibited the IL-1β induced degradation of I-

κBβ, thereby increasing NF-kB cytoplasmic retention. LTF did not inhibit I-κBβ mRNA induction as evidenced by its increased presence in cells treated with IL-1β and LTF.

This is not surprising as I-κBβ expression is not regulated by NF-κB transcription factors. Thus, the presence of I-κBβ may provide the necessary sequestration of the NF-kβ in the cytoplasm in the absence of I-κBα.

Upregulation of Runx2, Osteopontin and Alkaline phosphatase gene expression is the hallmark of osteogenic differentiation (127, 151). Alkaline phosphatase plays a critical role in the mineralization of the osteoid matrix, while Runx2 is a prerequisite for osteoblast differentiation (38, 127). Since areas exposed to low magnitudes of tensile forces exhibit bone formation clinically, (126, 148-150) we examined whether the signals generated by LTF regulate the expression of these genes. The catabolic signals in the presence of IL-1β inhibited the expression of these osteogenic markers and LTF abrogated this inhibition and upregulated the expression of all three genes to control levels. These

120

findings suggest that upregulation of osteogenic markers may yet be another functional mechanism by which LTF induces bone formation, as opposed to HTF in the presence of

IL-1β.

Above in vitro findings correlate with the in vivo observations and suggest that tensile forces of low magnitudes are likely to exert anti-inflammatory signals in the PDL subjected to mechanical loading. Similarly, tensile forces of high magnitudes induce expression of IL-1β and COX-2 in vivo and in vitro, providing evidence for the proinflammatory nature of tensile forces at higher magnitudes. The observed bone deposition in tensile fields could be the result of the potent anabolic nature of the signals generated by

LTF, similar to those observed in vitro. Through correlation, these findings provide a mechanism that may explain a role of mechanical signals in bone adaptation to mechanical loading. Furthermore, the study offers an explanation as to how different types of mechanical forces are perceived by the osteoblast-like cells leading to variations in their biologic responses.

ACKNOWLEDGEMENTS

The study was supported by grants from the National Institute of Health AR048781,

AT00646, and DEO15399.

121

FIGURE 13. LTF inhibits synthesis of pro-inflammatory mediators COX-2 (A) and IL-

1β (Β) in the rat PDL cells subjected to mechanical loading. Rat first molars were sub-

jected to 40 cN of orthodontic forces, the hemi-maxillae were fixed and immunostained

with anti-IL-1β and COX-2 antibodies. Areas of compression (C) and tension (T) in the

cervical and apical regions of the roots exhibit inflammation at both the compression and

tension sites in cervical region. The apical areas that received lower forces exhibit mini-

mal COX-2 and IL-1b expression at the tension sites but pronounced expression at the

compression sites. (C) A composite image of the hemi-maxilla from the cervical and

apical regions. Cervical section shows the expression of COX-2 at the compression and

tension sites around the first and second molars suggesting that the forces were experienced by both teeth. Apical section shows the expression of COX-2 at the compression sites of the first molar roots, but not at the sites exposed to low tensile forces. Third molars used as no touch controls show absence of any inflammation. (D) Diagrammatic representation of the tipping of first molars showing the areas experiencing compressive and tensile forces, as well as areas of high and low tensile forces. Data shown in A-C is representative one from three separate hemimaxillae

122

Figure 13

123

Figure 13 continued

124

FIGURE 14. LTF and HTF differentially regulate COX-2 and IL-1β mRNA expression in osteoblast like PDL cells. (A) PDL cells were either untreated, exposed to IL-1β alone, exposed to LTF alone, or exposed simultaneously to IL-1β and LTF for 24 hours. (B)

Cells were treated similar to shown in (A) except LTF was replaced by HTF. Subse- quently, RNA was extracted and the expression of IL-1β and COX-2 mRNA analyzed by real time PCR. The expression of GAPDH was used as an internal control to equalize the mRNA input. Data shown is from three independent experiments. * indicates p< 0.0

125

Figure 14

126

FIGURE 15. LTF inhibits the nuclear translocation, mRNA expression, and protein synthesis of NF-κB p65. (A) Immunofluorescence analysis of PDL cells demonstrating inhibition of IL-1β induced nuclear translocation of NF-κB P65 by LTF at 30, 60, and 120 minutes. Actin is stained with FITC-Phalloidin, and NF-κB is stained with rabbit anti

NF-κB p65 IgG and CY3 conjugated goat anti rabbit IgG as second antibody (red). (B)

NF-κB mRNA analysis exhibiting IL-1β induces upregulation of NF-κBp65 at 15, 30,

60 and 90 minutes, whereas concomitant exposure of LTF suppresses IL-1β induced NF-

κB p65 mRNA expression. (C) Inhibition of IL-1β induced synthesis of NF-κB P65 by

LTF. Graphs shows densitometric analysis of the bands representing NF-κB p65 in the

Western blots (panel in inset). Data shown is representative of one of three (A and B) and one of two independent experiments in (C).

127

Figure 15 128

FIGURE 16. Regulation of I-κB degradation and synthesis by LTF. (A) Densitometric analysis of Western blots (inset panel) showing IL-1β induces rapid degradation of I-κBα and I-κBβ at 15, 30, 60, 90 and 120 mins. LTF inhibits I-κBα protein degradation during first 30 min and its loss is observed at 60 to 120 min, whereas LTF inhibits I-κBβ degradation at all time points. (B) IL-1β induces the mRNA expression of I-κBα at 15, 30, 60,

90 and 120 mins. Simultaneous exposure of cells to LTF and IL-1β simultaneously inhibits IL-1β-dependent of I-κBα mRNA expression. Data shown is representative of one out of two experiments in (A) and one of three experiments in (B).

129

Figure 16

130

FIGURE 17. LTF acts as an anabolic signal. Real time PCR analysis showing IL-1β in-

duced inhibition of Runx2, osteopontin and alkaline phosphatase mRNA expression is

abrogated by LTF. Data represent mean and standard error from three independent experiments performed in duplicates. * p< 0.05 when cells treated with IL-1β and IL-1β

+LTF were compared

131

Figure17

132

CHAPTER 6

CONCLUSIONS

Many recent investigations have demonstrated, in vitro and in vivo, that both exercise and passive motion can exert reparative effects on inflamed joints, whereas inappropriate or excessive mechanical forces initiate cartilage destruction such as observed in osteoarthritic joints. However, the specific intracellular signaling cascades that transduce mechanical signals into the biochemical events responsible for cartilage destruction or repair remain somewhat elusive, sometimes paradoxical, and clearly complex. This disseratation summarizes how the signals generated by biomechanical stress may initiate either the repair or destruction of cartilage, depending upon their magnitude, frequency and duration.

While the use of subjectively “appropriate” joint movement has been suggested for centuries to be therapeutic for arthritic joints, the specific molecular mechanisms and defined biomechanical signals that regulate inflammation and pro-

133

mote cartilage repair remain an enigma. Examination of the transcriptional regulation of pro-inflammatory and matrix-associated genes has provided a framework for examining how mechanical signals regulate cartilage metabolism. Chondrocytes respond to all three major types of biomechanical forces, in magnitude-, frequency- and duration-dependent manners. However, it appears that dynamic forces provide the initiating stimulus capable of suppressing proinflammatory gene inductions, while static forces invariably induce these proinflammatory gene expressions. Dynamic biomechanical signals of high magnitudes activate pro-inflammatory genes via the activation of the NF-κB family of transcription factors. Conversely, at lower/physiological magnitudes these signals attenuate pro-inflammatory gene induction. The biomechanical signals generated by appropriate magnitudes of cyclic tensile strain are able to attenuate IL-1β-dependent activation of IKK and result in the inhibition of NF-κB transcriptional activity. These signals act at multiple steps within the NF-κB signaling cascade to inhibit the transcriptional activity of NF-κB itself, by pre- venting its nuclear import, as well as inhibiting the activation and gene expression of its inhibitors I-κBα and I-κBβ. Biomechanical signal-enforced dynamic shuttling of NF-κB/I-κBα complexes may represent yet another level of regulation, which could inhibit the activity of NF-κB and the subsequent downstream signaling events involved in pro-inflammatory gene induction. Compressive, tensile and shear forces of appropriate low/physiological magnitudes also promote the up-regulation of proteoglycan and collagen synthesis that is drasti-

134

cally inhibited in inflamed joints. Thus, the beneficial effects of physiological levels of biomechanical signals or exercise may be explained by their ability to suppress the signal transduction pathways of proinflammatory/catabolic mediators, while simultaneously stimulating the anabolic pathways. Whether these anabolic signals are a consequence of the inhibition of NF-κB signaling, or are mediated via distinct anabolic pathways has yet to be elucidated. Regardless, it is this complex interplay of magnitude- and frequency-dependent reparative signals within the inflammatory microenvironment of the cartilage that ultimately allows the remarkable and beneficial effects of dynamic tensile forces to be realized.

135

BIBLIOGRAPHY

1. Grodzinsky AJ, Levenston ME, Jin M, Frank EH. Cartilage tissue remodeling in response to mechanical forces. Annu.Rev.Biomed.Eng. 2000;2:691-713.

2. Park S, Hung CT, Ateshian GA. Mechanical response of bovine articular cartilage under dynamic unconfined compression loading at physiological stress levels. Osteoarthritis Cartilage 2004;12(1):65-73.

3. Fehrenbacher A, Steck E, Rickert M, Roth W, Richter W. Rapid regulation of collagen but not metalloproteinase 1, 3, 13, 14 and tissue inhibitor of metalloproteinase 1, 2, 3 expression in response to mechanical loading of cartilage explants in vitro. Arch.Biochem.Biophys. 2003;410(1):39-47.

4. Fitzgerald JB, Jin M, Grodzinsky AJ. Shear and compression differentially regulate clusters of functionally related temporal transcription patterns in cartilage tissue. J.Biol.Chem. 2006;281(34):24095-24103.

5. Giannoni P, Siegrist M, Hunziker EB, Wong M. The mechanosensitivity of cartilage oligomeric matrix protein (COMP). Biorheology 2003;40(1-3):101-109.

6. Murata M, Bonassar LJ, Wright M, Mankin HJ, Towle CA. A role for the interleukin-1 receptor in the pathway linking static mechanical compression to decreased proteoglycan synthesis in surface articular cartilage. Arch.Biochem.Biophys. 2003;413(2):229-235.

7. Valhmu WB, Raia FJ. myo-Inositol 1,4,5-trisphosphate and Ca(2+)/calmodulin- dependent factors mediate transduction of compression-induced signals in bovine articular chondrocytes. Biochem.J. 2002;361(Pt 3):689-696.

8. Valhmu WB, Stazzone EJ, Bachrach NM, Saed-Nejad F, Fischer SG, Mow VC, et al. Load-controlled compression of articular cartilage induces a transient stimulation of aggrecan gene expression. Arch.Biochem.Biophys. 1998;353(1):29-36.

9. Mio K, Saito S, Tomatsu T, Toyama Y. Intermittent compressive strain may reduce aggrecanase expression in cartilage: a study of chondrocytes in agarose gel. Clin.Orthop.Relat.Res. 2005;433:225-232.

136

10. Mauck RL, Byers BA, Yuan X, Tuan RS. Regulation of cartilaginous ECM gene transcription by chondrocytes and MSCs in 3D culture in response to dynamic loading. Biomech.Model.Mechanobiol 2007;6(1-2):113-125.

11. Kisiday J, Jin M, Kurz B, Hung H, Semino C, Zhang S, et al. Self-assembling peptide hydrogel fosters chondrocyte extracellular matrix production and cell division: implications for cartilage tissue repair. Proc.Natl.Acad.Sci.U.S.A. 2002;99(15):9996-10001.

12. Hunter CJ, Imler SM, Malaviya P, Nerem RM, Levenston ME. Mechanical compression alters gene expression and extracellular matrix synthesis by chondrocytes cultured in collagen I gels. Biomaterials 2002;23(4):1249-1259.

13. Stoddart MJ, Ettinger L, Hauselmann HJ. Enhanced matrix synthesis in de novo, scaffold free cartilage-like tissue subjected to compression and shear. Biotech- nol.Bioeng. 2006 Dec 20;95(6):1043-1051.

14. Grad S, Gogolewski S, Alini M, Wimmer MA. Effects of Simple and Complex Motion Patterns on Gene Expression of Chondrocytes Seeded in 3D Scaffolds. Tissue Eng. 2006 Oct 1.

15. Fitzgerald JB, Jin M, Dean D, Wood DJ, Zheng MH, Grodzinsky AJ. Mechanical compression of cartilage explants induces multiple time-dependent gene expression patterns and involves intracellular calcium and cyclic AMP. J.Biol.Chem. 2004 May 7;279(19):19502-19511.

16. Lee JH, Fitzgerald JB, Dimicco MA, Grodzinsky AJ. Mechanical injury of cartilage explants causes specific time-dependent changes in chondrocyte gene expression. Arthritis Rheum. 2005;52(8):2386-2395.

17. Chowdhury TT, Bader DL, Lee DA. Dynamic compression counteracts IL-1beta induced iNOS and COX-2 activity by human chondrocytes cultured in agarose constructs. Biorheology 2006;43(3-4):413-429.

18. Chowdhury TT, Bader DL, Lee DA. Dynamic compression counteracts IL-1 beta-induced release of nitric oxide and PGE2 by superficial zone chondrocytes cultured in agarose constructs. Osteoarthritis Cartilage 2003;11(9):688-696.

19. Chowdhury TT, Appleby RN, Salter DM, Bader DA, Lee DA. Integrin-mediated mechanotransduction in IL-1 beta stimulated chondrocytes. Bio- mech.Model.Mechanobiol 2006 Jun;5(2-3):192-201.

20. Chowdhury TT, Bader DL, Lee DA. Anti-inflammatory effects of IL-4 and dynamic compression in IL-1beta stimulated chondrocytes. Bio- chem.Biophys.Res.Commun. 2006;339(1):241-247.

137

21. Guilak F, Fermor B, Keefe FJ, Kraus VB, Olson SA, Pisetsky DS, et al. The role of biomechanics and inflammation in cartilage injury and repair. Clin.Orthop.Relat.Res. 2004; 423:17-26.

22. Upton ML, Chen J, Guilak F, Setton LA. Differential effects of static and dynamic compression on meniscal cell gene expression. J.Orthop.Res. 2003;21(6):963-969.

23. Tomiyama T, Fukuda K, Yamazaki K, Hashimoto K, Ueda H, Mori S, et al. Cy- clic compression loaded on cartilage explants enhances the production of reactive oxygen species. J.Rheumatol. 2007;34(3):556-562.

24. Griffin TM, Guilak F. The role of mechanical loading in the onset and progres- sion of osteoarthritis. Exerc.Sport Sci.Rev. 2005;33(4):195-200.

25. Alexopoulos LG, Setton LA, Guilak F. The biomechanical role of the chondrocyte pericellular matrix in articular cartilage. Acta Biomater. 2005;1(3):317-325.

26. Lucchinetti E, Bhargava MM, Torzilli PA. The effect of mechanical load on integrin subunits alpha5 and beta1 in chondrocytes from mature and immature cartilage explants. Cell Tissue Res. 2004;315(3):385-391.

27. Orazizadeh M, Cartlidge C, Wright MO, Millward-Sadler SJ, Nieman J, Halliday BP, et al. Mechanical responses and integrin associated protein expression by human ankle chondrocytes. Biorheology 2006;43(3-4):249-258.

28. Chowdhury TT, Salter DM, Bader DL, Lee DA. Integrin-mediated mechanotransduction processes in TGFbeta-stimulated monolayer-expanded chondrocytes. Biochem.Biophys.Res.Commun. 2004;318(4):873-881.

29. Guilak F. The deformation behavior and viscoelastic properties of chondrocytes in articular cartilage. Biorheology 2000;37(1-2):27-44.

30. Knight MM, Toyoda T, Lee DA, Bader DL. Mechanical compression and hydrostatic pressure induce reversible changes in actin cytoskeletal organisation in chondrocytes in agarose. J.Biomech. 2006;39(8):1547-1551.

31. Long P, Gassner R, Agarwal S. Tumor necrosis factor alpha-dependent proinflammatory gene induction is inhibited by cyclic tensile strain in articular chondrocytes in vitro. Arthritis Rheum. 2001;44(10):2311-2319.

32. Madhavan S, Anghelina M, Rath-Deschner B, Wypasek E, John A, Deschner J, et al. Biomechanical signals exert sustained attenuation of proinflammatory gene induction in articular chondrocytes. Osteoarthritis Cartilage 2006 May 25.

138

33. Deschner J, Hofman CR, Piesco NP, Agarwal S. Signal transduction by mechanical strain in chondrocytes. Curr.Opin.Clin.Nutr.Metab.Care 2003;6(3):289- 293.

34. Gassner R, Buckley MJ, Piesco N, Evans C, Agarwal S. Cytokine-induced nitric oxide production of joint cartilage cells in continuous passive movement. Anti- inflammatory effect of continuous passive movement on chondrocytes: in vitro study. Mund Kiefer Gesichtschir. 2000; Suppl 2:S479-84.

35. Gassner R, Buckley MJ, Georgescu H, Studer R, Stefanovich-Racic M, Piesco NP, et al. Cyclic tensile stress exerts antiinflammatory actions on chondrocytes by inhibiting inducible nitric oxide synthase. J.Immunol. 1999;163(4):2187- 2192.

36. Xu Z, Buckley MJ, Evans CH, Agarwal S. Cyclic tensile strain acts as an antagonist of IL-1 beta actions in chondrocytes. J.Immunol. 2000;165(1):453-460.

37. Ferretti M, Srinivasan A, Deschner J, Gassner R, Baliko F, Piesco N, et al. Anti- inflammatory effects of continuous passive motion on meniscal fibrocartilage. J.Orthop.Res. 2005;23(5):1165-1171.

38. Agarwal S, Deschner J, Long P, Verma A, Hofman C, Evans CH, et al. Role of NF-kappaB transcription factors in antiinflammatory and proinflammatory actions of mechanical signals. Arthritis Rheum. 2004;50(11):3541-3548.

39. Deschner J, Wypasek E, Ferretti M, Rath B, Anghelina M, Agarwal S. Regula- tion of RANKL by biomechanical loading in fibrochondrocytes of meniscus. J.Biomech. 2006;39(10):1796-1803.

40. Deschner J, Rath-Deschner B, Agarwal S. Regulation of matrix metalloproteinase expression by dynamic tensile strain in rat fibrochondrocytes. Os- teoarthritis Cartilage 2006;14(3):264-272.

41. Angele P, Yoo JU, Smith C, Mansour J, Jepsen KJ, Nerlich M, et al. Cyclic hydrostatic pressure enhances the chondrogenic phenotype of human mesenchymal progenitor cells differentiated in vitro. J.Orthop.Res. 2003;21(3):451-457.

42. Angele P, Schumann D, Angele M, Kinner B, Englert C, Hente R, et al. Cyclic, mechanical compression enhances chondrogenesis of mesenchymal progenitor cells in tissue engineering scaffolds. Biorheology 2004;41(3-4):335-346.

43. Trindade MC, Shida J, Ikenoue T, Lee MS, Lin EY, Yaszay B, et al. Intermittent hydrostatic pressure inhibits matrix metalloproteinase and pro-inflammatory mediator release from human osteoarthritic chondrocytes in vitro. Osteoarthritis Cartilage 2004;12(9):729-735.

139

44. Griffin BD, Moynagh PN. In vivo binding of NF-kappaB to the IkappaBbeta promoter is insufficient for transcriptional activation. Biochem.J. 2006;400(1):115-125.

45. Anghelina M, Dossumbekova A, Madhavan S, He L, Quan N, Knobloch T, Agarwal S. Biomechanical Signals Inhibit IKK Activity to Attenuate NF-κB Transcriptional Activity in Inflamed Chondrocytes. Arthritis and Rheumatism 2007 (In review)

46. Lane Smith R, Trindade MC, Ikenoue T, Mohtai M, Das P, Carter DR, et al. Ef- fects of shear stress on articular chondrocyte metabolism. Biorheology 2000;37(1-2):95-107.

47. Lee MS, Trindade MC, Ikenoue T, Goodman SB, Schurman DJ, Smith RL. Regulation of nitric oxide and bcl-2 expression by shear stress in human osteoarthritic chondrocytes in vitro. J.Cell.Biochem. 2003;90(1):80-86.

48. Healy ZR, Lee NH, Gao X, Goldring MB, Talalay P, Kensler TW, and Konstan- topoulos K. Divergent responses of chondrocytes and endothelial cells to shear stress: Cross-talk among COX-2, the phase 2 response, and apoptosis. PNAS 2005; 102:14010–14015.

49. Agarwal S, Long P, Gassner R, Piesco NP, and Buckley MJ. Cyclic tensile strain suppresses catabolic effects of interleukin-1β in fibrochondrocytes from the temporomandibular joint. Arthritis Rheum 44: 608–617, 2001.

50. Aufderheide AC and Athenesiou KA. Comparison of scaffolds and culture conditions for tissue engineering of the knee meniscus. Tissue Eng 11: 1095–1104, 2005.

51. Benjamin M and Hillen B. Mechanical influences on cells, tissues and organs: mechanical morphogenesis. Eur J Morphol 41: 3–7, 2003.

52. Brosseau L, Milne S, and Wells GA. Efficacy of continuous passive motion following total knee arthroplasy: a metaanalysis. J Rheumatol 31: 2251–2264, 2004.

53. Buckwalter JA. Sports, joint injury, and posttraumatic osteoarthritis. J Orthop Sports Phys Ther 33: 578–588, 2003.

54. Chowdhury TT, Bader DL, and Lee DA. Dynamic compression inhibits the synthesis of nitric oxide and PGE2 by IL-1 -stimulated chondrocytes cultured in agarose constructs. Biochem Biophys Res Commun 285: 1168–1174, 2001.

55. Evans CH, Robbins PD, Ghivizzani SC, Wasko MC, Tomaino MM, Kang R, Muzzonigro TA, Vogt M, Elder EM, Whiteside TL, Watkins SC, and Herndon JH. Gene transfer to human joints: progress toward a gene therapy of arthritis. Proc Natl Acad Sci USA 102: 8698–8703, 2005. 140

56. Fermor B, Jeffcoat D, Hennerbichler A, Pisetsky DS, Weinberg JB, and Guilak F. The effects of cyclic mechanical strain and tumor necrosis factor-_ on the response of cells of the meniscus. Osteoarthritis Cartilage 12: 956–962, 2004.

57. Fink C, Fermor B, Weinberg JB, Pisetsky DS, Misukonis MA, and Guilak F. The effect of dynamic mechanical compression on nitric oxide production in the meniscus. Osteoarthritis Cartilage 9: 481–487, 2001.

58. Gabrion A, Aimedieu P, Laya Z, Havet E, Mertl P, Grebe R, and Laude M. Re- lationship between ultrastructure and biomechanical properties of the knee meniscus. Surg Radiol Anat 27: 507–510, 2005.

59. Harada Y, Tomita N, Nakajima M, Ikeuchi K, and Wakitani S. Effect of low loading and joint immobilization for spontaneous repair of osteochondral defect in the knees of weightless (tail suspension) rats. J Orthop Sci 10: 508–514, 2005.

60. Iannone F and Lapadula G. The pathophysiology of osteoarthritis. Aging Clin Exp Res 15: 364–372, 2003.

61. Kim HK, Kerr RG, Cruz TF, and Salter RB. Effects of continuous passive motion and immobilization on synovitis and cartilage degradation in antigen induced arthritis. J Rheumatol 22: 1714–1721, 1995.

62. Krane SM and Goldring MB. Clinical implications of cartilage metabolism in arthritis. Eur J Rheumatol Inflamm 10: 4–9, 1990.

63. Lee DA, Frean SP, Lees P, and Bader DL. Dynamic mechanical compression influences nitric oxide production by articular chondrocytes seeded in agarose. Biochem Biophys Res Commun 251: 580–585, 1998.

64. Li KW, Wang AS, and Sah RL. Microenvironment regulation of extracellular signal-regulated kinase activity in chondrocytes: effects of culture configuration, interleukin-1, and compressive stress. Arthritis Rheum 48: 689–699, 2003.

65. Marsolais GS, Dvorak G, and Conzemius MG. Effects of postoperative rehabilitation on limb function after cranial cruciate ligament repair in dogs. J Am Vet Med Assoc 220: 1325–1330, 2002.

66. McDevitt CA and Webber RJ. The ultrastructure and biochemistry of meniscal cartilage. Clin Orthop Relat Res 252: 8–18, 1990.

67. Mow V and Chen H. Structure and function of articular cartilage and meniscus. In: Basic Orthopedic Biomechanics and Mechanobiology, edited by Mow VC and Huisckes R (3rd ed.). New York: Lippincott and Williams, 2005.

68. Patwari P, Cook MN, DiMicco MA, Blake SM, James IE, Kumar S, Cole AA, Lark MW, and Grodzinsky AJ. Proteoglycan degradation after injurious com- 141

pression of bovine and human articular cartilage in vitro: interaction with exogenous cytokines. Arthritis Rheum 48: 1292–1301, 2003.

69. Shin S, Fermor B, Weinberg BB, Pisetsky DS, and Guilak F. Regulation of matrix turnover in meniscal explants: role of mechanical stress, interleukin-1, and nitric oxide. J Appl Physiol 95: 308–313, 2003.

70. Studer RK. Nitric oxide decreases IGF-1 receptor function in vitro; glutathione depletion enhances this effect in vivo. Osteoarthritis Cartilage12: 863–869, 2004.

71. Sweigart MA and Athanasiou KA. Toward tissue engineering of the knee meniscus. Tiss Eng 7: 111–129, 2001.

72. Sweigert MA, Zhu CF, Burt DM, deHall PD, Agarwal CM, Clanton TO, and Athanasiou KA. Intraspecies and interspecies comparison of the compressive properties of the medial meniscus. Ann Biomed Eng 11: 1569–1579, 2004.

73. Szafranski JD, Grodzinsky AJ, Burger E, Gaschen V, Hung HH, and Hunziker EB. Chondrocyte mechanotransduction: effects of compression on deformation of intracellular organelles and relevance to cellular biosynthesis. Osteoarthritis Cartilage 12: 937–946, 2004.

74. Buckwalter JA, Mankin HJ, Grodzinsky AJ. Articular cartilage and osteoarthritis. Instr Course Lect 2005; 54:465e80.

75. Kurz B, Lemke AK, Fay J, Pufe T, Grodzinsky AJ, Schunke M. Pathomecha- nisms of cartilage destruction by mechanical injury. Ann Anat 2005;187:473e85.

76. Milne S, Brosseau L, Robinson V, Noel MJ, Davis J, Drouin H, et al. Continuous passive motion following total knee arthroplasty. Cochrane Database Syst Rev 2003;2:CD004260.

77. Das UN. Anti-inflammatory nature of exercise. Nutrition 2004;20:323e6.

78. Bennell K, Hinman R. Exercise as a treatment for osteoarthritis. Curr Opin Rheumatol 2005;17:634e40.

79. Ostrowski K, Rohde T, Asp S, Schjerling P, Pedersen BK. Pro- and anti- inflammatory cytokine balance in strenuous exercise in humans. J Physiol 1999; 515:287e91.

80. Rannou F, Poiraudeau S, Revel M. Cartilage: from biomechanics to physical therapy. Ann Readapt Med Phys 2001;44:259e67.

81. Tran-Khanh N, Hoemann CD, McKee MD, Henderson JE, Buschmann MD. Aged bovine chondrocytes display a diminished capacity to produce a collagen-

142

rich, mechanically functional cartilage extracellular matrix. J Orthop Res 2005;23:1354e62.

82. Roos EM, Dahlberg L. Positive effects of moderate exercise on glycosaminogly- can content in knee cartilage: a four-month, randomized, controlled trial in patients at risk of osteoarthritis. Arthritis Rheum 2005;52:3507e14.

83. Piscoya JL, Fermor B, Kraus VB, Stabler TV, Guilak F. The influence of mechanical compression on the induction of osteoarthritis-related biomarkers in articular cartilage explants. Osteoarthritis Cartilage 2005;13:1092e9.

84. Dequeker J, Dieppe PA. Disorders of bone cartilage and connective tissue. In: Klippel JH, Dieppe PA, Eds. Rheumatology. 2nd edn. London: Mosby 1998.

85. Pelletier JP, DiBattista JA, Roughley P, McCollum R, Martel-Pelletier J. Cyto- kines and inflammation in cartilage degradation. J Rheum Dis Clin North Am 1993; 19:545e68.

86. Mazzetti I, Magagnoli G, Paoletti S, Uguccioni M, Olivotto E, Vitellozzi R, et al. A role for chemokines in the induction of chondrocyte phenotype modulation. Arhritis Rheum 2004;50:112e22.

87. Lapadula G, Iannone F, Zuccaro C, Grattagliano V, Covelli M. Chondrocyte phenotyping in human osteoarthritis. Clin Rheumatol 1998;17:99e104.

88. Kolettas E, Muir HI, Barrett J, Hardingham TE. Chondrocyte phenotype and cell survival are regulated by culture conditions and by specific cytokines through the expression of Sox-9 transcription factor. Rheumatology 2001;40:1146e56.

89. Fermor B, Weinberg JB, Pisetsky DS, Misukonis MA, Fink C, Guilak F. Induc- tion of cyclooxygenase-2 by mechanical stress through a nitric oxide-regulated pathway. Osteoarthritis Cartilage 2002;10:792e8.

90. Karjalainen HM, Sironen RK, Elo MA, Kaarniranta K, Takigawa M, Helminen HJ. Gene expression profiles in chondrosarcoma cells subjected to cyclic stretch- ing and hydrostatic pressure. A cDNA array study. Biorheology 2003;40:93e100.

91. Long P, Gassner R, Agarwal S. Tumor necrosis factor alpha-dependent proinflammatory gene induction is inhibited by cyclic tensile strain in articular chondrocytes in vitro. Arthritis Rheum 2001;44:2311e9.

92. Tew SR, Li Y, Pothacharoen P, Tweats LM, Hawkins RE, Hardingham TE. Ret- roviral transduction with SOX9 enhances re-expression of the chondrocyte phenotype in passaged osteoarthritic human articular chondrocytes. Osteoarthritis Cartilage 2005; 13:80e9.

143

93. Burrage PS, Mix KS, Brinckerhoff CE. Matrix metalloproteinases: role in arthritis. Front Biosci 2006;11: 529e43.

94. Studer RK, Jaffurs D, Stefanovic-Racic M, Robbins PD,Evans CH. Nitric oxide in osteoarthritis. Osteoarthritis Cartilage 1999;7:377e9.

95. Hardy MM, Seibert K, Manning PT, Currie MG, Woerner BM, Edwards D, et al. Cyclooxygenase 2-dependent prostaglandin E2 modulates cartilage proteoglycan degradation in human osteoarthritis explants. Arthritis Rheum 2002;46:1789e803.

96. Evans CH, Watkins SC, Stefanovic-Racic M. Nitric oxide and cartilage metabolism. Methods Enzymol 1996;269:7e13.

97. Murrell GA, Jang D, Williams RJ. Nitric oxide activates metalloprotease in articular cartilage. Biochem Biophys Res Commun 1995;206:15.

98. Lianxu, C., J. Hongti, and Y. Changlong. 2006. NF-kappaBp65-specific siRNA inhibits expression of genes of COX-2, NOS-2 and MMP-9 in rat IL-1beta- induced and TNF-alpha-induced chondrocytes. Osteoarthritis Cartilage 14: 367- 376

99. Hauselmann HJ, Stefanovic-Racic M, Michel BA, Evans CH. Differences in nitric oxide production by superficialand deep human articular chondrocytes: implications for proteoglycan turnover in inflammatory joint diseases. J Immunol 1998;160:1444e8.

100. Watanabe H, de Caestecker MP, Yamada Y. Transcriptional cross-talk between Smad, ERK1/2, and p38 mitogen-activated protein kinase pathways regulates transforming growth factor-b-induced aggrecan gene expression in chondrogenic ATDC5 cells. J Biol Chem 2001;276:14466e73.

101. Gaffo, A., K. G. Saag, and J. R. Curtis. 2006. Treatment of rheumatoid arthritis. Am. J. Health. Syst. Pharm. 63: 2451-2465.

102. Abramson, S. B. and Y. Yazici. 2006. Biologics in development for rheumatoid arthritis: relevance to osteoarthritis. Adv. Drug Deliv. Rev. 58: 212-225.

103. Sivalingam, S. P., J. Thumboo, S. Vasoo, S. T. Thio, C. Tse, and K. Y. Fong. 2007. In vivo Pro- and Anti-inflammatory Cytokines in Normal and Patients with Rheumatoid Arthritis. Ann. Acad. Med. Singapore 36: 96-94.

104. Weissmann, G. 2006. The pathogenesis of rheumatoid arthritis. Bull. Hosp. Jt. Dis. 64: 12-15.

105. Moreland, L. W. 2004. Biologic therapies on the horizon for rheumatoid arthritis. J. Clin. Rheumatol. 10: S32-9. 144

106. Schopf, L., A. Savinainen, K. Anderson, J. Kujawa, M. DuPont, M. Silva, E. Siebert, S. Chandra, J. Morgan, P. Gangurde, D. Wen, J. Lane, Y. Xu, M. Hep- perle, G. Harriman, T. Ocain, and B. Jaffee. 2006. IKKbeta inhibition protects against bone and cartilage destruction in a rat model of rheumatoid arthritis. Ar- thritis Rheum. 54: 3163-3173.

107. Tas, S. W., J. Adriaansen, N. Hajji, A. C. Bakker, G. S. Firestein, M. J. Ver- voordeldonk, and P. P. Tak. 2006. Amelioration of arthritis by intraarticular dominant negative Ikk beta gene therapy using adeno-associated virus type 5. Hum. Gene Ther. 17: 821-832.

108. Clohisy, J. C., B. C. Roy, C. Biondo, E. Frazier, D. Willis, S. L. Teitelbaum, and Y. Abu-Amer. 2003. Direct inhibition of NF-kappa B blocks bone erosion associated with inflammatory arthritis. J. Immunol. 171: 5547-5553.

109. Ferretti, M., R. Gassner, Z. Wang, P. Perera, J. Deschner, G. Sowa, R. B. Salter, and S. Agarwal. 2006. Biomechanical signals suppress proinflammatory responses in cartilage: early events in experimental antigen-induced arthritis. J. Immunol. 177: 8757-8766.

110. Ateshian, G. A. 2007. Artificial cartilage: weaving in three dimensions. Nat. Mater. 6: 89-90.

111. Garcia, A. J. and D. Boettiger. 1999. Integrin-fibronectin interactions at the cell- material interface: initial integrin binding and signaling. Biomaterials 20: 2427- 2433.

112. Goldmann, W. H. 2002. Mechanical aspects of cell shape regulation and signaling. Cell Biol. Int. 26: 313-317.

113. Hsieh, M. H. and H. T. Nguyen. 2005. Molecular mechanism of apoptosis induced by mechanical forces. Int. Rev. Cytol. 245: 45-90.

114. Ingber, D. E. 1997. Tensegrity: the architectural basis of cellular mechanotransduction. Annu. Rev. Physiol. 59: 575-599.

115. Gassner, R. J., M. J. Buckley, R. K. Studer, C. H. Evans, and S. Agarwal. 2000. Interaction of strain and interleukin-1 in articular cartilage: effects on proteoglycan synthesis in chondrocytes. Int. J. Oral Maxillofac. Surg. 29: 389-394.

116. Shim, J. H., C. Xiao, A. E. Paschal, S. T. Bailey, P. Rao, M. S. Hayden, K. Y. Lee, C. Bussey, M. Steckel, N. Tanaka, G. Yamada, S. Akira, K. Matsumoto, and S. Ghosh. 2005. TAK1, but not TAB1 or TAB2, plays an essential role in multiple signaling pathways in vivo. Genes Dev. 19: 2668-2681.

117. DiDonato, J. A. 2000. Assaying for I kappa B kinase activity. Methods Enzy- mol. 322: 393-400. 145

118. Hoffmann, A., A. Levchenko, M. L. Scott, and D. Baltimore. 2002. The Ikap- paB-NF-kappaB signaling module: temporal control and selective gene activation. Science 298: 1241-1245.

119. Hayden, M. S. and S. Ghosh. 2004. Signaling to NF-kappaB. Genes Dev. 18: 2195-2224.

120. Senftleben, U. and M. Karin. 2002. The IKK/NF-kappa B pathway. Crit. Care Med. 30: S18-26.

121. Griffin, B. D. and P. N. Moynagh. 2006. Persistent interleukin-1beta signaling causes long term activation of NFkappaB in a promoter-specific manner in human glial cells. J. Biol. Chem. 281: 10316-10326.

122. Moynagh, P. N. 2005. The NF-kappaB pathway. J. Cell. Sci. 118: 4589-4592.

123. Buss, H., A. Dorrie, M. L. Schmitz, R. Frank, M. Livingstone, K. Resch, and M. Kracht. 2004. Phosphorylation of serine 468 by GSK-3beta negatively regulates basal p65 NF-kappaB activity. J. Biol. Chem. 279: 49571-49574.

124. Pomerantz, J. L. and D. Baltimore. 2002. Two pathways to NF-kappaB. Mol. Cell 10: 693-695.

125. Sindelar BJ, Herring SW. 2005. Soft tissue mechanics of the temporomandibular joint. Cells Tissues Organs 180:36-43.

126. Davidovitch Z 1991 Tooth movement. Crit Rev Oral Biol Med 2:411-450.

127. Meikle MC 2006 The tissue, cellular, and molecular regulation of orthodontic tooth movement: 100 years after Carl Sandstedt. Eur J Orthod 28:221-240.

128. 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 17:1545-1554.

129. Warden SJ, Hurst JA, Sanders MS, Turner CH, Burr DB, Li J 2005 Bone adaptation to a mechanical loading program significantly increases skeletal fatigue re- sistance. J Bone Miner Res 20:809-816.

130. Gluhak-Heinrich J, Pavlin D, Yang W, MacDougall M, Harris SE 2007 MEPE expression in osteocytes during orthodontic tooth movement. Arch Oral Biol 52:684-690.

131. Raab-Cullen DM, Thiede MA, Petersen DN, Kimmel DB, Recker RR 1994 Mechanical loading stimulates rapid changes in periosteal gene expression. Cal- cif Tissue Int 55:473-478.

146

132. Patel MJ, Liu W, Sykes MC, Ward NE, Risin SA, Risin D, Jo H 2007 Identifi- cation of mechanosensitive genes in osteoblasts by comparative microarray studies using the rotating wall vessel and the random positioning machine. J Cell Biochem 101:587-599.

133. Rubin J, Rubin C, Jacobs CR 2006 Molecular pathways mediating mechanical signaling in bone. Gene 367:1-16.

134. Fan X, Roy E, Zhu L, Murphy TC, Ackert-Bicknell C, Hart CM, Rosen C, Nanes MS, Rubin J 2004 Nitric oxide regulates receptor activator of nuclear factor-kappaB ligand and osteoprotegerin expression in bone marrow stromal cells. Endocrinology 145:751-759.

135. Cherian PP, Siller-Jackson AJ, Gu S, Wang X, Bonewald LF, Sprague E, Jiang JX 2005 Mechanical strain opens connexin 43 hemichannels in osteocytes: a novel mechanism for the release of prostaglandin. Mol Biol Cell 16:3100-3106.

136. Krishnan V, Davidovitch Z 2006 The effect of drugs on orthodontic tooth movement. Orthod Craniofac Res 9:163-171.

137. de Carlos F, Cobo J, Diaz-Esnal B, Arguelles J, Vijande M, Costales M 2006 Orthodontic tooth movement after inhibition of cyclooxygenase-2. Am J Orthod Dentofacial Orthop 129:402-406.

138. Basaran G, Ozer T, Kaya FA, Kaplan A, Hamamci O 2006 Interleukine-1beta and tumor necrosis factor-alpha levels in the human gingival sulcus during orthodontic treatment. Angle Orthod 76:830-836.

139. Bletsa A, Berggreen E, Brudvik P 2006 Interleukin-1alpha and tumor necrosis factor-alpha expression during the early phases of orthodontic tooth movement in rats. Eur J Oral Sci 114:423-429.

140. Long P, Hu J, Piesco N, Buckley M, Agarwal S 2001 Low magnitude of tensile strain inhibits IL-1beta-dependent induction of pro-inflammatory cytokines and induces synthesis of IL-10 in human periodontal ligament cells in vitro. J Dent Res 80:1416-1420.

141. Long P, Liu F, Piesco NP, Kapur R, Agarwal S 2002 Signaling by mechanical strain involves transcriptional regulation of proinflammatory genes in human periodontal ligament cells in vitro. Bone 30:547-552.

142. Kawarizadeh A, Bourauel C, Zhang D, Gotz W, Jager A 2004 Correlation of stress and strain profiles and the distribution of osteoclastic cells induced by orthodontic loading in rat. Eur J Oral Sci 112:140-147.

147

143. McGuinness NJ, Wilson AN, Jones ML, Middleton J 1991 A stress analysis of the periodontal ligament under various orthodontic loadings. Eur J Orthod 13:231-242.

144. Liu ZJ, King GJ, Gu GM, Shin JY, Stewart DR 2005 Does human relaxin accelerate orthodontic tooth movement in rats? Ann N Y Acad Sci 1041:388-394.

145. Madan MS, Liu ZJ, Gu GM, King GL 2007 Effects of human relaxin on orthodontic tooth movement and periodontal ligaments in rats. Am J Orthod Dentafa- cial Orthop 131:8.e1-8.e10

146. Agarwal S, Long P, Seyedain A, Piesco N, Shree A, Gassner R 2003 A central role for the nuclear factor-kappaB pathway in anti-inflammatory and proinflammatory actions of mechanical strain. FASEB J 17:899-901.

147. Ghosh S, Karin M 2002 Missing pieces in the NF-kappaB puzzle. Cell 109 Suppl:S81-96.

148. Isaacson RJ, Lindauer SJ, Davidovitch M 1993 On tooth movement. Angle Or- thod 63:305-309.

149. Masella RS, Meister M 2006 Current concepts in the biology of orthodontic tooth movement. Am J Orthod Dentofacial Orthop 129:458-468.

150. Stanfeld J, Jones J, Laster L, Davidovitch Z 1986 Biochemical aspects of orthodontic tooth movement. I. Cyclic nucleotide and prostaglandin concentrations in tissues surrounding orthodontically treated teeth in vivo. Am J Orthod Dentofa- cial Orthop 90:139-148.

151. Holleville N, Mateos S, Bontoux M, Bollerot K, Monsoro-Burq AH 2007 Dlx5 drives Runx2 expression and osteogenic differentiation in developing cranial su- ture mesenchyme. Dev Biol 304:860-874.

148