University of IowaIowa Research Online

Theses and Dissertations

Spring 2011

The isoprenoid biosynthesis pathway andregulation of osteoblast differentiationMegan Moore WeivodaUniversity of Iowa

Copyright 2011 Megan Moore Weivoda

This dissertation is available at Iowa Research Online: https://ir.uiowa.edu/etd/1106

Follow this and additional works at: https://ir.uiowa.edu/etd

Part of the Pharmacology Commons

Recommended CitationWeivoda, Megan Moore. "The isoprenoid biosynthesis pathway and regulation of osteoblast differentiation." PhD (Doctor ofPhilosophy) thesis, University of Iowa, 2011.https://doi.org/10.17077/etd.fycowap2

THE ISOPRENOID BIOSYNTHESIS PATHWAY AND REGULATION OF OSTEOBLAST DIFFERENTIATION

by

Megan Moore Weivoda

An Abstract

Of a thesis submitted in partial fulfillment of the requirements for the Doctor of Philosophy

degree in Pharmacology in the Graduate College of The University of Iowa

May 2011

Thesis Supervisor: Professor Raymond J. Hohl

1

Statins, drugs commonly used to lower serum cholesterol, have been shown to

stimulate osteoblast differentiation and bone formation. By inhibiting HMG-CoA

reductase (HMGCR) statins deplete the cellular isoprenoid biosynthetic pathway products

farnesyl pyrophosphate (FPP) and geranylgeranyl pyrophosphate (GGPP). Current

thought in the field is that statins stimulate bone formation through the depletion of

GGPP, since exogenous GGPP prevents the effects of statins on osteoblasts in vitro.

We hypothesized that direct inhibition of GGPP synthase (GGPPS) would

similarly stimulate osteoblast differentiation. Digeranyl bisphosphonate (DGBP), a

specific inhibitor of GGPPS, decreased GGPP levels in MC3T3-E1 pre-osteoblasts and

calvarial osteoblasts leading to impaired protein geranylgeranylation. In contrast to our

hypothesis, DGBP inhibited the matrix mineralization of MC3T3-E1 cells and the

expression of osteoblast differentiation markers in calvarial osteoblasts. The effect on

mineralization was not prevented by exogenous GGPP. By inhibiting GGPPS, DGBP led

to an accumulation of the GGPPS substrate FPP. We show that FPP and GGPP levels

decreased during MC3T3-E1 and calvarial osteoblast differentiation, which correlated

with decreased expression of HMGCR and FPP synthase. The decrease in FPP during

differentiation was prevented by DGBP treatment. The accumulation of FPP following 24

h DGBP treatment correlated with activation of the glucocorticoid receptor, suggesting a

potential mechanism by which DGBP-induced FPP accumulation may inhibit osteoblast

differentiation.

To further investigate whether FPP inhibits osteoblast differentiation, we utilized

the squalene synthase (SQS) inhibitor zaragozic acid (ZGA), which causes a greater

accumulation of FPP than can be achieved with GGPPS inhibition. ZGA treatment

2

decreased osteoblast proliferation, gene expression, alkaline phosphatase (ALP) activity,

and matrix mineralization of calvarial osteoblasts. Prevention of ZGA-induced FPP

accumulation with HMGCR inhibition prevented the effects of ZGA on osteoblast

differentiation. Treatment of osteoblasts with exogenous FPP similarly inhibited matrix

mineralization. These results suggest that the accumulation of FPP negatively regulates

osteoblast differentiation.

While we did not find that specific depletion of GGPP stimulates osteoblast

differentiation, we obtained evidence that GGPP does negatively regulate the

differentiation of these cells. Exogenous GGPP treatment inhibited primary calvarial

osteoblast gene expression and matrix mineralization. Interestingly, GGPP pre-treatment

increased markers of insulin signaling, despite reduced phosphorylation of the insulin

receptor (InsR). Inhibition of osteoblast differentiation by GGPP led to the induction of

PPARγ and enhanced adipogenesis in osteoblastic cultures, suggesting that GGPP may

play a role in the osteoblast versus adipocyte fate decision. Adipogenic differentiation of

primary bone marrow stromal cell (BMSC) cultures was prevented by DGBP treatment.

Altogether these data present novel roles for the isoprenoids FPP and GGPP in the

regulation of osteoblast differentiation and have intriguing implications for the isoprenoid

biosynthetic pathway in the regulation of skeletal homeostasis.

Abstract Approved: ______________________________________________ Thesis Supervisor ______________________________________________ Title and Department ______________________________________________ Date

THE ISOPRENOID BIOSYNTHESIS PATHWAY AND REGULATION OF OSTEOBLAST DIFFERENTIATION

by

Megan Moore Weivoda

A thesis submitted in partial fulfillment of the requirements for the Doctor of Philosophy

degree in Pharmacology in the Graduate College of The University of Iowa

May 2011

Thesis Supervisor: Professor Raymond J. Hohl

Graduate College The University of Iowa

Iowa City, Iowa

CERTIFICATE OF APPROVAL

____________________________

PH.D. THESIS

_____________

This is to certify that the Ph. D. thesis of

Megan Moore Weivoda

has been approved by the Examining Committee for the thesis requirement for the Doctor of Philosophy degree in Pharmacology at the May 2011 graduation.

Thesis Committee: ________________________________ Raymond J. Hohl, Thesis Supervisor ________________________________ Christopher M. Adams ________________________________ John G. Koland ________________________________ Frederick W. Quelle ________________________________ David F. Wiemer

For the LORD gives wisdom; From His mouth come knowledge and understanding;

Proverbs 2:6 The Bible

ii

ACKNOWLEDGEMENTS

I would like to begin by extending my extreme gratitude to my thesis advisor, Dr.

Raymond J. Hohl, for welcoming me into his lab and allowing me the opportunity to

investigate the isoprenoid pathway in osteoblast differentiation. Additionally, I want to

thank my thesis committee members: Drs. Christopher M. Adams, John G. Koland,

Frederick W. Quelle, and David F. Wiemer for their time and guidance. I also wish to

thank the Department of Pharmacology for allowing me the opportunity to join their

department and continue my academic studies.

I would also like to thank past and present members of the Hohl laboratory.

Specifically, thanks to Dr. Amel Dudakovic and Brian Wasko for the lunches and coffee

breaks that led to stimulating scientific discussions. Also, thanks to Dr. Craig Kuder for

his advice and helpful encouragement throughout my time at Iowa. I would also like to

thank Dr. Huaxiang Tong for his help in measuring intracellular isoprenoids.

I would like to thank the Vanderbilt Center for Bone Biology. My time there laid

the foundation for my love of bone biology. Specifically, thanks to Drs. James Edwards

and XiangLi Yang Elefteriou for instruction in the proper study of bone biology. I would

like to extend my gratitude to the late Dr. Gregory R. Mundy (1942-2010). His

enthusiasm for bone biology and scientific discovery will live on in me as I continue my

academic career.

I would like to thank my family as well as my husband’s family for all of their

love and support. To my parents, Larry and Cindy Moore, thanks for always believing in

me. And last, but not least, I wish to thank my husband, Aaron Weivoda, for his constant

love, encouragement, and support of my scientific career.

iii

iv

TABLE OF CONTENTS

LIST OF TABLES vi LIST OF FIGURES vii LIST OF ABBREVIATIONS ix CHAPTER I: INTRODUCTION 1

The Skeleton and Bone Remodeling 1 Osteoporosis 2 Isoprenoid Biosynthetic Pathway 5 Osteoblast Differentiation and the Isoprenoid Biosynthetic Pathway 8 Hypothesis 12

CHAPTER II: INHIBITION OF GERANYLGERANYL PYROPHOSPHATE SYNTHASE (GGPPS) AND OSTEOBLAST DIFFERENTIATION 16

Abstract 16 Introduction 17 Materials and Methods 18 Results 21 Discussion 24

CHAPTER III: ACCUMULATION OF FPP AND OSTEOBLAST DIFFERENTIATION 34

Abstract 34 Introduction 35 Materials and Methods 36 Results 40 Discussion 44

CHAPTER IV: GGPP AND THE OSTEOBLAST VERSUS ADIPOGENIC FATE DECISION 58

Abstract 58 Introduction 59 Materials and Methods 61 Results 66 Discussion 72

CHAPTER V: SUMMARY 88

The Isoprenoid Pathway and Osteoblast Differentiation 88 Future Studies 93 Conclusion 95

REFERENCES 96

v

LIST OF TABLES

Table 1. Rat primers for real-time qPCR analysis 29

vi

LIST OF FIGURES

Figure 1. Skeletal bone remodeling 14 Figure 2. The isoprenoid biosynthetic pathway 15 Figure 3. DGBP reduces intracellular GGPP and impairs protein

geranylgeranylation 30 Figure 4. DGBP inhibits osteoblast differentiation and matrix

mineralization 31 Figure 5. DGBP leads to an accumulation of intracellular FPP 32 Figure 6. DGBP leads to activation of the glucocorticoid receptor 33 Figure 7. The isoprenoid biosynthetic pathway and primary

calvarial osteoblast differentiation 48 Figure 8. ZGA leads to increases in endogenous FPP and GGPP 49 Figure 9. ZGA leads to inhibition of osteoblast differentiation 50 Figure 10. SQSI-154 causes an accumulation of FPP and inhibits

osteoblast matrix mineralization 52 Figure 11. Inhibition of HMGCR prevents FPP accumulation and

inhibition of osteoblast differentiation in cells treated with ZGA 53

Figure 12. GGPP depletion does not restore osteoblast

mineralization 55 Figure 13. Addition of exogenous mevalonate and co-treatment with

FTI-277 56 Figure 14. GGPP inhibits primary calvarial osteoblast differentiation 78 Figure 15. GGPP increases the expression of PPARγ 79 Figure 16. Inhibition of PPARγ transcriptional activity does not

prevent the effects of GGPP on osteoblast mineralization 80 Figure 17. GGPP enhances adipogenesis 81

vii

Figure 18. Specific inhibition of GGPPS inhibits adipogenesis 83 Figure 19. GGPP treatment does not increase geranylgeranylation 84 Figure 20. GGPP reduces InsR phosphorylation 85 Figure 21. GGPP enhances insulin-induced Erk1/2 activation and

glucose uptake 86 Figure 22. Proposed mechanisms for the inhibition of osteoblast

differentiation by GGPP 87

viii

LIST OF ABBREVIATIONS

ALP Alkaline Phosphatase

BMD Bone Mineral Density

BMP Bone Morphogenetic Protein

BMSC Bone Marrow Stromal Cell

BP Bisphosphonate

C/EBPα CCAAT/Enhancer-Binding Protein α

Col1a1 Type I Collagen

DGBP Digeranyl Bisphosphonate

DMAPP Dimethylallyl Pyrophosphate

FOH Farnesol

FPP Farnesyl Pyrophosphate

FPPS Farnesyl Pyrophosphate Synthase

FTase Farnesyl Transferase

FTI Farnesyl Transferase Inhibitor

GGOH Geranylgeraniol

GGPP Geranylgeranyl Pyrophosphate

GGPPS Geranylgeranyl Pyrophosphate Synthase

GGTase Geranylgeranyl Transferase

Glut4 Glucose Transporter 4

GPP Geranyl Pyrophosphate

GR Glucocorticoid Receptor

HAP Hydroxyapatite

HMGCR HMG-CoA Reductase

IGF-1 Insulin-like Growth Factor 1

ix

x

InsR Insulin Receptor

IPP Isopentenyl Pyrophosphate

IRS-1 Insulin Receptor Substrate

LDL Low-density Lipoprotein

M-CSF Macrophage Colony Stimulating Factor

MSC Mesenchymal Stem Cell

NBP Nitrogenous Bisphosphonate

OCN Osteocalcin

OPG Osteoprotegrin

OST-PTP Osteotesticular Protein Tyrosine Phosphatase

PPAR Peroxisome Proliferator-Activated Receptor

PTH Parathyroid Hormone

PTPase Protein Tyrosine Phosphatase

RANKL Receptor Activator of NFkB Ligand

Runx2 Runt Related Transcription Factor 2

SNP Single Nucleotide Polymorphism

SQS Squalene Synthase

TGFβ Transforming Growth Factor beta

UCP1 Uncoupling Protein 1

ZGA Zaragozic Acid

1

CHAPTER I: INTRODUCTION

The Skeleton and Bone Remodeling

The skeleton serves several vital functions in human physiology, including

mechanical support, protection of internal organs, storage and metabolism of calcium and

phosphate, and regulation of hematopoiesis(1, 2). Because of these important roles, it is

necessary that there is a mechanism to replace microfractures and weakened bone that

result from daily stress on the skeleton. This process is referred to as bone remodeling (1,

3, 4).

Bone remodeling occurs throughout the adult lifespan and consists of bone

resorption by multi-nucleated, hematopoietic-derived osteoclasts and bone formation by

mesenchymal-derived osteoblasts. Bone resorption and formation are balanced so that the

amount of bone resorbed is equal to the amount of new bone formed (1-3). This balance

is controlled by intricate signaling pathways that determine the replication,

differentiation, function, and death of bone resorbing and forming cells(3). Much of this

signaling occurs directly between the osteoblasts and osteoclasts, otherwise referred to as

osteoblast-osteoclast coupling(5). Osteoblasts secrete factors that stimulate and inhibit

osteoclast differentiation and activity. Receptor activator of NFκB ligand (RANKL) and

macrophage colony stimulating factor (M-CSF) are produced by osteoblasts and are

necessary for osteoclastic differentiation from their hematopoietic precursors.

Osteoprotegerin (OPG), which is also produced by osteoblasts, acts as a soluble receptor

for RANKL, preventing the association of RANKL with RANK on pre-cursor or mature

osteoclasts. Osteoblasts control the ratio of secreted RANKL and OPG to regulate

2

osteoclastic differentiation and activity(1, 4). In turn, osteoclasts influence osteoblast

differentiation and bone formation through the resorption-dependent release of insulin-

like growth factor (IGF)-1, transforming growth factor (TGF)-ß, and bone morphogenetic

proteins (BMPs) from the bone matrix. Release of these growth factors promote

osteoblast migration to resorption sites and stimulate osteoblast differentiation(1) (Figure

1).

Osteoporosis

The balance between bone resorption and bone formation is influenced by

genetic, nutritional, hormonal, and environmental factors(1, 6), and disruption of this

balance results in bone disease. Overactivity of osteoblasts (osteosclerosis) and decreased

osteoclast activity (osteopetrosis) result in high bone mass phenotypes due to bone

formation being in excess of bone resorption (1). In contrast, when bone resorption is in

excess of bone formation, low bone mass phenotypes occur. Osteoporosis is the most

common disease of low bone mass and is characterized by reduced bone mineral density

(BMD), leading to increased bone fragility and increased risk for fractures(1, 7, 8).

Osteoporosis can result from increased osteoclastic bone resorption due to the estrogen

deficiency following menopause (post-menopausal or type I osteoporosis) or the age-

related decrease in osteoblast number and vitality (senile osteoporosis or type II

osteoporosis)(1, 9).

Osteoporosis affects approximately eight million women and two million men in

the United States(3, 4). An additional 34 million individuals have osteopenia, or low bone

mass, putting them at risk for developing the disease(4). It is estimated that one in two

3

women and one in four men over the age of 55 will develop an osteoporotic fracture in

their lifetime(4). Worldwide there are approximately nine million osteoporotic fractures

each year, with over half of these occurring in Europe and the Americas(8). These

fractures are major health problems for the elderly, causing significant morbidity and

mortality(7, 8, 10). Osteoporotic fractures also result in high costs to health care services.

The direct costs of osteoporotic fractures in the United States in 2005 were estimated to

be $19 billion(4, 8, 10).

Current therapies for osteoporosis, including estrogen-replacement therapy

(ERT), selective estrogen receptor modifiers (SERMs), calcitonin, denosumab, and the

bisphosphonates (BPs), aim to inhibit osteoclastic bone resorption. ERT and SERMs act

to replace or mimic the estrogen that is lost in the post-menopausal period and that

contributes to osteoporosis. Activation of the estrogen receptor in osteoblasts increases

the ratio of secreted OPG to RANKL resulting in decreased osteoclast differentiation and

bone resorption (4, 9). Calcitonin is a hormone that binds to the calcitonin receptor on

osteoclasts to inhibit resorptive activity (4, 11). Denosumab is a monoclonal antibody

targeting RANKL that has recently been approved by the United States Food and Drug

Administration (FDA). By acting as a soluble receptor for RANKL, similar to OPG,

Denosumab prevents the differentiation and activity of osteoclasts(4, 5, 11). Currently,

the most widely prescribed drugs for osteoporosis are the BPs (4, 11, 12).

The BPs are synthetic analogs of pyrophosphate, consisting of a non-hydrolyzable

phosphonate-carbon-phosphonate (P-C-P) backbone with two variable side chains, R1

and R2. The BP core gives these compounds a strong affinity for bone mineral (5, 12).

This affinity can be further enhanced by a hydroxyl group in the R1 position and a

4

nitrogen group in the R2 position. Nitrogen containing BPs are referred to as nitrogenous

BPs (NBPs), whereas BPs without a nitrogen in this position are referred to as non-

nitrogenous BPs (non-NBPs). BPs are released from the bone surface by the acidic pH

generated in the osteoclast resorption lacunae, allowing uptake of the drug into the

osteoclasts (12). Non-NBPs act by being incorporated into adenosine triphosphate (ATP),

resulting in the formation of non-hydrolyzable ATP analogues and osteoclast apoptosis.

In contrast, the more potent NBPs inhibit the enzyme farnesyl pyrophosphate synthase

(FPPS) as depicted in Figure 2(13-15). As will be discussed in greater detail later, FPPS

is an enzyme in the isoprenoid biosynthetic pathway. Inhibition of FPPS leads to

decreased levels of the isoprenoids farnesyl pyrophosphate (FPP) and geranylgeranyl

pyrophosphate (GGPP). These molecules are necessary for the isoprenylation and

membrane localization of several proteins involved in the osteoclastic F-actin ring

formation, and the subsequent development of the resorption lacunae. Inhibition of FPPS

by the NBPs prevents the development of the F-actin ring and resorption lacunae, leading

to the inhibition of osteoclast activity. There is also some speculation that NBPs may

inhibit the differentiation of pre-osteoclasts into mature osteoclasts (5, 12). By inhibiting

FPPS, NBPs also lead to an accumulation of the FPPS substrate isopentenyl

pyrophosphate (IPP). Monkkonen, et al. demonstrated that this accumulation of IPP leads

to the formation of a novel ATP analogue, ApppI. ApppI inhibits the mitochondrial

ADP/ATP translocase and results in osteoclast apoptosis(16). However preliminary

reports describing mice deficient for bim, a protein involved in apoptosis, suggest that the

main mechanism for inhibition of osteoclastic bone resorption by NBPs is inhibition of

protein prenylation and osteoclast resorptive activity (17). By preventing osteoclastic

5

bone resorption via inhibition of osteoclast differentiation, stimulation of osteoclast

apoptosis, or inhibition of osteoclast resorptive activity, BPs reduce the risk of

osteoporotic fractures (4, 11). NBPs, such as zoledronate, have also proven useful in the

treatment of osteolytic bone diseases, such as multiple myeloma and metastatic breast

cancer, acting to prevent the tumor-induced osteoclastic bone resorption; these drugs may

also have direct effects on tumor cells(18).

While anti-resorptive therapies are effective in preventing further bone loss and

reducing the incidence of fractures, they are not able to restore bone structure that has

been lost (3). Anabolic agents that stimulate osteoblastic bone formation have the

potential to reconstruct the skeleton and to restore bone structure. This would be

especially beneficial for patients who have sustained substantial bone loss (19). Currently

the only anabolic agent approved by the FDA is the recombinant parathyroid hormone

(PTH) peptide fragment, teriparatide. Teriparatide is administered intermittently and

results in an increase in the number and activity of osteoblasts. However, because of its

high cost, teriparatide use is currently confined to only high risk populations (1).

Therefore, much research is being done to develop novel agents to stimulate osteoblast

differentiation and bone formation as a treatment for osteoporosis and other conditions of

low bone mass.

Isoprenoid Biosynthetic Pathway

The isoprenoid biosynthetic pathway is the source of numerous biological

compounds (Figure 2). β-Hydroxymethylglutaryl coenzyme A (HMG-CoA), derived

from acetyl-CoA, is reduced to mevalonate in a reaction catalyzed by HMG-CoA

6

reductase (HMGCR). Mevalonate is then phosphorylated by mevalonate kinase, to form

5-phosphomevalonate. This compound undergoes an additional phosphorylation and

decarboxylation to form the five-carbon isopentenyl pyrophosphate (IPP), the basic

building block referred to as an isoprene unit. IPP isomerizes to dimethylallyl

pyrophosphate (DMAPP), and together IPP and DMAPP are converted to the ten-carbon

geranyl pyrophosphate (GPP) by FPPS. Addition of a second molecule of IPP to GPP by

FPPS results in the formation of the fifteen-carbon FPP. FPP is the branch point of the

isoprenoid biosynthetic pathway. Squalene synthase catalyzes the head to head

condensation of two FPP molecules to form squalene, the first committed step towards

cholesterol biosynthesis. Alternatively, GGPP synthase (GGPPS) combines FPP and IPP

to form the twenty-carbon GGPP. GGPP is a precursor for several longer chain

isoprenoids, including ubiquinone (20, 21).

In addition to serving as substrates for the synthesis of sterols and longer-chain

non-sterol isoprenoids, FPP and GGPP are utilized in the isoprenylation of proteins.

Isoprenylation is a posttranslational protein modification and consists of the addition of

an isoprenoid group to a carboxy-terminal cysteine. This modification is necessary for the

correct localization of certain proteins, including small GTPases, and can therefore

influence protein activity as well as protein-protein interactions(21-23). In a search of the

Swiss-Prot database, 300 proteins were identified as potential targets of protein

prenylation (24). FPP is the substrate for farnesyl transferase (FTase) and GGPP is the

substrate for geranylgeranyl transferases I and II (GGTases I and II), resulting in

farnesylation and geranylgeranylation of proteins, respectively. FTase and GGTase I are

heterodimeric, zinc metalloenzymes made up of a common α subunit and divergent ß

7

subunits. FTase and GGTase I recognize the carboxy-terminal CAAX motif, in which C

is the prenylated cysteine residue and A represents aliphatic residues. The presence of

serine, methionine, alanine, or glutamine at the X residue designates the protein as an

FTase substrate, whereas a leucine designates the protein as a GGTase I substrate.

Following this prenylation, the AAX peptide is cleaved, and the cysteine residue is

methylated. Similar to FTase and GGTase I, GGTase II, otherwise known as Rab

GGTase, is a heterodimeric protein made up of GGTase II α and ß subunits. In contrast to

FTase and GGTase I, GGTase II requires Rab escort protein 1 (Rep1) for its activity.

GGTase II targets CC, CAC, CCX, or CCXX carboxy-terminal motifs and

geranylgeranylates one or both cysteine residues (21-23).

FPP and GGPP, as well as their metabolites, have also been shown to act as

ligands for certain receptors. The alcohol form of FPP, faresol (FOH) activates the

farnesoid X receptor (FXR)(25) whereas the alcohol form of GGPP, geranylgeraniol

(GGOH) has been shown to inhibit the activity of liver X receptor (LXR)-α(26). More

recently, Das, et al. demonstrated that FPP binds and activates the estrogen, thyroid, and

glucocorticoid receptors (27). Interestingly, the activation of the glucocorticoid receptor

by FPP occurred in primary keratinocytes and resulted in inhibition of wound healing in

skin organ cultures (28). In addition to activating nuclear hormone receptors, FPP has

been found to be an agonist of the G-protein coupled receptor (GPCR), GPR92 (29), and

an antagonist of the LPA3 receptor (30). It has also been demonstrated that FPP activates

the calcium channel TRPv3 resulting in nociceptive behaviors in inflamed animals,

suggesting that FPP is a novel endogenous pain-producing substance (31).

8

There are several FDA-approved pharmaceutical agents that target enzymes in the

isoprenoid biosynthetic pathway, namely the statins and NBPs. Statins are commonly

used to lower serum cholesterol (32) and act by inhibiting HMG-CoA reductase (33). By

inhibiting HMG-CoA reductase, statins inhibit cholesterol biosynthesis, leading to an

upregulation of the low-density lipoprotein (LDL) receptor. LDL receptor upregulation

leads to the increased clearance of plasma LDL cholesterol. In addition to depletion of

cholesterol, the inhibition of HMG-CoA reductase by statins leads to decreased

production of the isoprenoids FPP and GGPP. As described above, FPP and GGPP have

several biological functions and their depletion by statins is thought to give rise to

pleiotropic effects, such as the activation of endothelial nitric oxide synthase (eNOS)

leading to blood vessel dilation(32, 34).

NBPs make up a second class of pharmaceutically available agents targeting the

isoprenoid biosynthetic pathway. These drugs, as mentioned previously, are potent

inhibitors of osteoclastic bone resorption and are widely used in the management of

osteoporosis and cancer-induced osteolytic bone disease(12). The pyrophosphate-like

backbone targets these agents to the bone mineral, where they are taken up by the

osteoclast. NBPs inhibit FPPS leading to the depletion of the isoprenoids FPP and GGPP.

This leads to the impaired protein prenylation of several small GTPases that are

necessary for osteoclast activity, and therefore inhibition of bone resorption(13-15).

Osteoblast Differentiation and the Isoprenoid Biosynthesis Pathway

Osteoblasts are derived from mesenchymal stem cells (MSCs) in the bone marrow

(1). MSCs are multi-potential cells capable of differentiating into adipocytes, myoblasts,

9

fibroblasts, chondrocytes and osteoblasts (1, 2). Differentiation into these various cell

types is controlled by the expression of lineage-specific transcription factors (2, 35). In

differentiating into mature osteoblasts, MSCs commit to the osteo-chondroprogenitor

lineage, followed by the osteoprogenitor lineage. These cells then undergo an expansion

followed by maturation into mature osteoblasts expressing osteocalcin (OCN) and type I

collagen (Col1a1). Importantly, these cells exhibit the ability to form collagen-based

extracellular matrices and mineral deposits (1, 35).

As mentioned above, statins deplete FPP and GGPP leading to several pleiotropic

effects(32). Mundy, et al. demonstrated that statins stimulate osteoblast differentiation

and bone formation through the induction of BMP-2 expression (36). These results have

been subsequently supported by several in vitro(37-44) and in vivo(45, 46) studies. There

have also been positive correlations cited between higher BMDs and patients

administered lipophilic statins(47-50). The ability of statins to stimulate osteoblast

differentiation has been attributed to the depletion of GGPP, since the addition of

exogenous GGPP or GGOH prevents the effects of statins in vitro (41, 43). Consistent

with this observation, NBPs, which similarly deplete endogenous GGPP, stimulate

osteoblast differentiation in in vitro MSC cultures (51, 52). It is important to note,

however, that some studies have found negative effects for NBPs on osteoblast

differentiation (53, 54). In support of GGPP depletion stimulating osteoblast

differentiation, a report by Yoshida, et al. demonstrated that the expression of GGPPS

decreases during the osteoblastic differentiation of the MC3T3-E1 pre-osteoblast cell

line(55). Consistent with this, Takase, et al. demonstrated that mevalonate kinase, the

enzyme downstream of HMGCR, decreases following treatment with the anabolic

10

peptide hormone, PTH(56). Addition of exogenous GGOH to MC3T3-E1 pre-osteoblast

cultures prevented osteoblast differentiation as measured by matrix mineralization and

the expression of the osteoblastic genes alkaline phosphatase (ALP) and BMP-2 (55).

In studying the differentiation of osteoblasts and bone formation, much attention

is paid to the differentiation of MSCs into an alternate cell type involved in bone

homeostasis, the adipocyte (2). Although initially assumed to be inert, adipocytes are

now known to be central players in energy homeostasis and insulin sensitivity (2, 57),

synthesizing and transporting lipids, as well as secreting adipokines, such as leptin and

adiponectin (2, 58).

The decreased bone mass characteristic of osteopenia and osteoporosis is

associated with an increase in bone marrow adiposity. Interestingly, factors that promote

osteoblastogenesis inhibit adipogenesis and vice versa (2, 6, 58). For example,

peroxisome proliferator-activated receptor (PPAR)-γ2, a member of the PPAR subfamily

of nuclear hormone receptors (57), is a positive regulator of adipogenic differentiation

and a negative regulator of osteoblast differentiation, acting to suppress the activity of the

osteoblast transcription factor runt-related transcription factor (Runx)-2 (2, 35, 58). As

mentioned previously, GGPP depletion by statin drugs stimulates osteoblast

differentiation(36). Consistent with stimulators of osteoblastogenesis inhibiting

adipogenesis, statin-mediated GGPP depletion results in the inhibition of adipogenic

differentiation, demonstrating the importance of GGPP and geranylgeranylation to

adipogenesis (42, 59). Interestingly, in contrast to the decreased GGPPS exhibited by

differentiating MC3T3-E1 pre-osteoblasts(55), GGPPS is highly expressed in adipose

11

tissue of the leptin-deficient ob/ob mice, and its expression increases during the

adipogenic differentiation of the 3T3-L1 cell line (60).

In addition to studying the role of the isoprenoid pathway in osteoblast

differentiation models through the use of small molecule inhibitors, there has been much

recent investigation into single nucleotide polymorphisms (SNPs) found in isoprenoid

pathway enzymes that may contribute to skeletal disease. Several studies have noted that

a non-coding SNP in the FPPS gene (rs2297480) may be a genetic marker for lower

BMD in postmenopausal Caucasian women. BMD was lower at all skeletal sites

measured in women with CC or CA genotypes as compared to the AA genotype at this

site (61). A separate study investigating this SNP demonstrated that individuals with the

CC genotype showed a decreased response to two years of treatment with NBPs(62).

Although the effect of this SNP on enzyme activity and FPP production has not been

studied, one would speculate that this non-coding SNP may lead to greater expression of

FPPS, making patients less responsive to NBP therapy. A SNP in GGPPS has also been

noted to have skeletal effects. In a recent study by Choi, et al. patients homozygous for a

deletion SNP at -8188A (rs3840452) in this enzyme displayed higher femoral neck BMD

and a decreased response to NBP therapy, however their population size was very small

(n=144) and separate studies will be required to confirm their results(63). If the

significance of this SNP on the skeleton is confirmed, one may speculate that these

patients exhibit lower GGPPS expression, leading to lower GGPP levels, resulting in

higher BMD. While the authors of these studies emphasized the impact of these SNPs on

osteoclastic activity, it cannot be ruled out that these SNPs may impact osteoblastic bone

12

formation, especially considering the negative effects of exogenous GGOH on

osteoblastic cultures published by Yoshida, et al(55).

Hypothesis

The body of evidence summarized above raises several questions. If statins lead

to osteoblast differentiation through the depletion of GGPP and subsequent impaired

geranylgeranylation, would specific inhibition of GGPPS similarly yield increased

osteoblast differentiation? Additionally, if one assumes that in standard conditions, all

geranylgeranylated proteins are isoprenylated, why does the addition of exogenous

GGOH inhibit osteoblast differentiation? And what is the impact of other isoprenoids?

Does FPP, which has been shown to activate numerous nuclear hormone receptors

involved in maintenance of the skeleton, influence osteoblast activity? The studies

presented in this manuscript examine these questions and aim to determine the influence

of the isoprenoid pathway on osteoblast differentiation.

Herein I document studies intended to address these questions. Digeranyl

bisphosphonate (DGBP), a novel inhibitor of GGPPS (Figure 2) developed by our and

Dr. David F. Wiemer’s lab at the University of Iowa(64, 65), was utilized to determine

whether specific depletion of GGPP led to enhanced osteoblast differentiation, similar to

statin-induced osteoblast differentiation. By inhibiting GGPPS, DGBP leads to the

accumulation of the substrate FPP. DGBP and squalene synthase (SQS) inhibitors, which

similarly cause FPP accumulation, were employed to determine whether FPP plays a

negative role in osteoblast differentiation. Lastly, studies were performed to determine

whether GGPP itself may influence the osteoblast versus adipocyte fate decision and to

13

determine potential mechanisms by which GGPP inhibits osteoblast differentiation. The

overall aim of this work was to test the hypothesis that the intermediates of the isoprenoid

biosynthetic pathway negatively regulate osteoblast differentiation.

14

Figure 1. Skeletal bone remodeling.

15

Figure 2. The isoprenoid biosynthetic pathway. This illustration diagrams the conversion of mevalonate to the isoprenoid metabolites (IPP, DMAPP, GPP, FPP, squalene, and GGPP). Specific enzymes responsible for production of certain metabolites are indicated in italics. Specific inhibitors of these enzymes are also noted.

16

CHAPTER II: THE EFFECTS OF DIRECTION INHIBITION OF GERANYLGERANYL PYROPHOSPHATE SYNTHASE (GGPPS) ON

OSTEOBLAST DIFFERENTIATION Abstract

Statins, drugs commonly used to lower serum cholesterol, have been shown to stimulate

osteoblast differentiation and bone formation. These effects have been attributed to the

depletion of geranylgeranyl pyrophosphate (GGPP). In this study, we tested whether

specific inhibition of GGPP synthase (GGPPS) with digeranyl bisphosphonate (DGBP)

would similarly lead to increased osteoblast differentiation. DGBP concentration

dependently decreased intracellular GGPP levels in MC3T3-E1 pre-osteoblasts and

primary rat calvarial osteoblasts, leading to impaired Rap1a geranylgeranylation. In

contrast to our hypothesis, 1 µM DGBP inhibited matrix mineralization in the MC3T3-E1

pre-osteoblasts. Consistent with this finding, DGBP inhibited the expression of alkaline

phosphatase (ALP) and osteocalcin (OCN) in primary osteoblasts. By inhibiting GGPPS,

DGBP caused an accumulation of the GGPPS substrate farnesyl pyrophosphate (FPP).

This effect was observed throughout the time course of MC3T3-E1 pre-osteoblast

differentiation. Interestingly, DGBP treatment led to activation of the glucocorticoid

receptor in MC3T3-E1 pre-osteoblast cells, consistent with recent findings that FPP

activates nuclear hormone receptors. These findings demonstrate that direct inhibition of

GGPPS, and the resulting specific depletion of GGPP, does not stimulate osteoblast

differentiation. This suggests that in addition to depletion of GGPP, statin-stimulated

osteoblast differentiation may depend on the depletion of upstream isoprenoids, including

FPP.

17

Introduction

As discussed previously, osteoporosis is a condition characterized by low BMD

that puts one at greater risk for debilitating fractures(7, 8), and current treatments inhibit

bone resorption by osteoclasts to prevent further bone loss and reduce fracture risk(3, 11).

However, anabolic agents that stimulate bone formation by osteoblasts are needed to

restore bone mass and structure in patients who have sustained substantial bone loss(3,

19).

Statins, drugs commonly used to lower serum cholesterol, have positive effects on

osteoblast differentiation and bone formation both in vitro(36, 37, 39, 41, 43) and in

vivo(36, 66, 67). Additionally, clinical trials have indicated a positive correlations

between increased BMD and patients administered lipophilic statins(47, 48).

Through inhibition of HMGCR, statins deplete downstream isoprenoid

metabolites, including FPP and GGPP, resulting in impaired protein isoprenylation. It is

thought that the effects of statins on bone are due to the depletion of GGPP, since

addition of exogenous GGPP or GGOH to statin-treated osteoblasts prevents the effects

of statins on osteoblast differentiation and matrix mineralization(41, 43). Consistent with

this hypothesis, NBPs, which inhibit FPPS(14, 15) and deplete cells of GGPP, have been

shown to stimulate osteoblast differentiation in vitro(51, 52) and prevent the negative

effects of dexamethasone on osteoblast differentiation(68). It is important to note,

however, that several other studies have reported that NBPs decrease osteoblast

proliferation and matrix mineralization(53, 54).

DGBP is a BP that specifically inhibits GGPPS(64, 65). We hypothesized that if

statin-stimulated osteoblast differentiation occurred through the depletion of GGPP,

18

direct inhibition of GGPPS with DGBP would similarly result in increased osteoblast

differentiation and matrix mineralization.

Materials and Methods

Cell culture. The MC3T3-E1 pre-osteoblast cell line was obtained from ATCC.

The animal protocol used for isolation of primary rat calvarial cells was approved by the

Institutional Animal Care and Use Committee at the University of Iowa. Primary rat

osteoblast cells were obtained by three sequential enzyme digestions of calvariae from

two day old neonatal Sprague-Dawley rats (Harlan). Digestions were performed with

0.05% collagenase type I and 1% trypsin (Invitrogen) in serum free α modified essential

medium (α-MEM) (Invitrogen) at 37 oC with shaking. The first two digestions (10 and

20 minutes, respectively) were discarded. Following the final digestion (60 minutes),

cells were centrifuged and resuspended in α-MEM containing 10% fetal bovine serum

(FBS) and 1x penicillin-streptomycin (Invitrogen). Cells were grown in a humidified

atmosphere with 5% CO2 at 37 oC in 10 cm plates.

Cells were subcultured for experiments at a density of 1x104 cells/cm2. All

experiments were carried out in osteoblast differentiation medium. This consisted of -

MEM with 10 mM β-glycerophosphate and 50 µg/mL L-ascorbic acid (Fischer

Scientific). Treatments were replaced every 3-4 days until the termination of the

experiment.

FPP/GGPP Quantification. MC3T3-E1 pre-osteoblasts and primary calvarial

osteoblasts were plated in 10 cm plates. Upon confluency cells were treated for with

DGBP (Wiemer lab, University of Iowa) (0.1, 1, 10, 100 µM) for 24 h. Alternatively,

19

MC3T3-E1 pre-osteoblasts were treated with or without 1µM DGBP for 1, 2, or 4 weeks.

FPP and GGPP levels were determined as previously reported by reverse phase

HPLC(69). Briefly, cells were washed twice with phosphate buffered saline (PBS)

(Invitrogen) and isoprenoid pyrophosphates were extracted from cells and used as

substrates for incorporation into fluorescent GCVLS or GCVLL peptides by

farnesyltransferase or geranylgeranyl transferase I. The prenylated fluorescent peptides

were separated by reverse phase HPLC and quantified by fluorescence detection. Total

FPP and GGPP levels were normalized to total protein content as measured by the

bichinconic (BCA) assay. Values are expressed as pmoles per mg protein.

Western blotting. MC3T3-E1 pre-osteoblasts were plated in 10 cm plates. Upon

confluency, cells were treated for 24 h. At the end of the experiment, media were

removed and cells were washed twice in PBS. Cells were collected with a cell scraper

after the addition of 2% sodium dodecyl sulfate (SDS)-lysis buffer. Lysates were

transferred to a 1.5 ml tube, heated to 100 oC for five minutes, and passed through a 27-

guage needle. Lysates were then centrifuged and supernatant transferred to a fresh 1.5 ml

tube. Protein concentrations were determined by the BCA method. Proteins were

resolved on 12% (Rap1a and Ras) or 7.5% (glucocorticoid receptor or Sp1) SDS-PAGE

gels and transferred to polyvinylidene difluoride membranes by electrophoresis. Primary

and secondary antibodies were added sequentially for 1 h and proteins were visualized

using an enhanced chemiluminescence detection kit from GE Healthcare. Anti pan-Ras

was obtained from InterBiotechnology. Rap1a, glucocorticoid receptor, Sp1, αTub, and

ßTub antibodies were obtained from Santa Cruz Biotechnology, Inc. Phospho (serine

211) glucocorticoid receptor antibody was obtained from Abcam. Horseradish

20

peroxidase-conjugated anti-mouse and anti-rabbit secondary IgG antibodies were from

GE Healthcare while anti-goat secondary IgG antibody was from Santa Cruz

Biotechnology, Inc.

Mineralization assay. MC3T3-E1 pre-osteoblasts were plated in 24-well plates.

Treatment began when cells reached confluency. After 28 days of culture, cells were

fixed for one hour in ice-cold 70% ethanol. Cells were then washed thoroughly with

deionized water; mineralization was detected with 40 mM Alizarin red, pH=4.2 for 15

minutes. Following staining, cells were washed thoroughly with deionized water to

remove non-specific Alizarin red. Images were captured using a Canon EOS Rebel XS.

Mineralization was quantified as described previously(70). Briefly, Alizarin Red was

eluted with 10% acetic acid per well. Plates were shaken on an orbital rotator for 15 min.

The acetic acid and cells were transferred to a 1.5 mL tube and heated to 85 oC for 10

minutes. The samples were then cooled on ice and centrifuged. The supernatant was

transferred to a new tube and 10% ammonium hydroxide was added. Samples were

vortexed vigorously and aliquots were transferred to a 96-well plate. The absorbance was

read at 405 nm on a Thermomax Microplate Reader (Molecular Devices). Absorbance

was compared to a standard curve; values were expressed as total mol Alizarin Red per

well.

Real-time quantitative polymerase chain reaction (qPCR). Primary calvarial

osteoblasts were plated in 6-well plates. Treatment was applied when cells reached

confluency. Eight days following the onset of treatment, total RNA was isolated from

each well using Qiashredders and the RNeasy Mini Kit (Qiagen). During the isolation, a

DNase step was performed (Qiagen). One µg of RNA was reverse-transcribed into cDNA

21

using the iScript cDNA Synthesis Kit (Bio-Rad). Real-time PCR was performed with

Sybr Green Master Mix (Applied Biosystems) with an ABI SDS 7900 HT instrument

(Applied Biosystems). The real-time protocol consisted of 2 min at 50oC, 10 min at

95oC, followed by 40 cycles of 95oC (15 sec) and 60oC (one min). Dissociation curves

were obtained following real-time qPCR to ensure the proper amplification of target

cDNAs. Primers were obtained from Integrated DNA Technologies (Iowa City, IA) and

eluted in TE Buffer (Ambion). Sequences and amplicon lengths are found in the Table 1.

Nuclear/Cytosol Fractionation. MC3T3-E1 pre-osteoblasts were plated in 10 cm

plates. Upon confluency, cells were treated with 1 µM dexamethasone (Sigma) or DGBP

(10, 50, and 100 µM). Cells were treated for 24 h. Nuclear and cytosolic fractions were

obtained with the nuclear/cytosol fractionation kit from Biovision. Cells were lysed and

fractionated as described by the kit protocol. Alternatively, whole cell lysate was

obtained by lysing the cells with RIPA buffer with added protease and phosphatase

inhibitors. Protein concentrations were determined with the BCA method. Western

blotting was performed as described above.

Statistical Analysis. Data are expressed as means +/- standard error of the mean.

All experiments were repeated at least twice with similar results. Differences between

two groups were compared using unpaired student’s t tests. Statistical significance was

defined by p values less than 0.05.

Results

DGBP decreases GGPP levels and impairs protein geranylgeranylation. To

determine whether DGBP inhibited GGPPS in MC3T3-E1 pre-osteoblasts and primary

22

rat calvarial osteoblasts, cells were treated with increasing concentrations of DGBP for

24 h. DGBP reduced intracellular GGPP levels at concentrations of 1-100 µM (Figure

3A). This decrease in GGPP correlated with the appearance of unprenylated Rap1a, a

geranylgeranylated protein, in MC3T3-E1 pre-osteoblasts (Figure 3B) and primary

calvarial osteoblasts (data not shown). Lovastatin, which reduces both FPP and GGPP,

impaired both the farnesylation of Ras (appearance of an upper band) and the

geranylgeranylation of Rap1a. In contrast, DGBP, which does not deplete FPP, had no

effect on the farnesylation of Ras. Impaired geranylgeranylation of Rap1a by lovastatin

and DGBP was prevented by the addition of 20 µM GGPP. However, impaired

farnesylation of Ras by lovastatin-mediated FPP depletion was not prevented by the

addition of exogenous GGPP.

Mineralization is inhibited by DGBP independently of GGPP depletion. To

determine the effects of depletion of GGPP on osteoblast mineralization, MC3T3-E1 pre-

osteoblasts were treated with DGBP for 28 days. One micromolar DGBP significantly

reduced osteoblast mineralization as evidenced by reduced Alizarin red staining as

determined visually (Figure 4A) or by elution and quantification of the stain by

absorbance measurement (Figure 4B). DGBP at concentrations of 1.0 µM or less had no

significant effect on osteoblast viability as measured by MTT assay throughout the time

course of osteoblast differentiation (data not shown).

To determine whether the negative effect of DGBP on osteoblast mineralization

was due to the depletion of GGPP, GGPP add-backs were performed. 20 µM GGPP

alone significantly reduced mineralization of MC3T3 cell cultures as compared to control

23

wells. The addition of GGPP with DGBP did not prevent the effects of DGBP on

osteoblast mineralization.

Expression of osteoblastic genes is inhibited by DGBP. In order to determine the

effects of GGPPS inhibition on osteoblast differentiation, primary rat calvarial

osteoblasts were treated with 1.0 M DGBP for eight days, followed by the assessment

of osteoblastic gene expression. As shown in Figure 4C, treatment of the osteoblast cells

for eight days with 1.0 µM DGBP significantly inhibited the expression of ALP and

OCN, with no effect on the expression of Col1a1.

DGBP leads to an accumulation of FPP. To determine whether the inhibition of

GGPPS perturbs the level of its substrate, FPP, intracellular FPP was measured in

MC3T3-E1 pre-osteoblasts and primary rat calvarial osteoblasts after 24 hours of DGBP

treatment. Treatment with DGBP resulted in a concentration-dependent increase in

intracellular FPP in both cell types (Figure 5A). To determine whether the effect of

DGBP on GGPP depletion (Figure 3A) and FPP accumulation (Figure 5A) was transient,

intracellular GGPP and FPP were assessed throughout the four-week time course of

MC3T3-E1 cell differentiation. As shown in Figure 5B, MC3T3-E1 cell GGPP levels

were significantly decreased after four weeks of control differentiation. One M DGBP

significantly decreased GGPP levels at each time point assessed. Similarly to GGPP,

MC3T3-E1 differentiation cultures exhibited decreased intracellular FPP after four weeks

of control differentiation. Treatment of the differentiation cultures with 1.0 M DGBP

resulted in a significant accumulation of FPP at each time point, and prevented the

decrease of FPP exhibited during control differentiation.

24

DGBP treatment leads to glucocorticoid receptor activation in osteoblasts. It has

been shown recently that certain nuclear hormones, including the thyroid receptor,

estrogen receptor, and glucocorticoid receptor (GR), are bound and activated by FPP (27,

28). To determine whether the accumulation of intracellular FPP by DGBP correlates

with activation of nuclear hormone receptors in osteoblasts, phosphorylation of the GR

was assessed in MC3T3-E1 pre-osteoblast cells following 24 hours of treatment with

DGBP (10, 50, or 100 µM) or 1.0 µM of the GR agonist dexamethasone. As shown in

Figure 6A, dexamethasone led to the appearance of phosphorylated (Serine 211)-GR.

Similar to treatment with the GR agonist, DGBP led to increased phosphorylated-GR

with all concentrations tested. The level of phosphorylated-GR normalized to total GR

was quantified using densitometry (Figure 6B).

Nuclear accumulation of the GR following DGBP treatment of the MC3T3-E1

pre-osteoblasts was also assessed. One micromolar dexamethasone led to a decrease in

cytosolic and an increase in nuclear GR (Figure 6C). DGBP treatment had no effect on

cytosolic levels of the GR, however, the higher concentrations of DGBP (50 and 100

µM) led to increased nuclear GR. The nuclear and cytosolic GR levels were quantified

using densitometry (Figure 6D).

Discussion

These data demonstrate that the specific inhibition of GGPPS in osteoblasts by the

BP DGBP results in the depletion of intracellular GGPP and impaired protein

geranylgeranylation. In contrast to our hypothesis, specific depletion of GGPP does not

25

lead to increased osteoblast differentiation. Importantly, inhibition of GGPPS leads to an

accumulation of the GGPPS substrate, FPP, and the subsequent activation of the GR.

As mentioned earlier, statins have been shown to stimulate osteoblast

differentiation and bone formation(36, 37, 39, 41, 43). These effects have been attributed

to decreased GGPP, since the addition of exogenous GGPP prevents statin-stimulated

osteoblast mineralization(41) and BMP-2 expression(43). Consistent with this, NBPs,

which similarly decrease GGPP, have been shown to stimulate osteoblast differentiation

in vitro(51, 52).

The results presented here have important consequences for the studies showing

positive effects of statins and NBPs on osteoblast differentiation. Direct inhibition of

GGPPS and depletion of GGPP inhibited osteoblast differentiation and matrix

mineralization of MC3T3-E1 pre-osteoblasts and primary rat calvarial osteoblasts. These

results suggest that the depletion of isoprenoids upstream of GGPP may be essential for

normal osteoblast differentiation.

One possibility is that DGBP inhibits osteoblast mineralization due to its

bisphosphonate core backbone. It has been reported that bisphosphonates display a

pyrophosphate-like effect to prevent hydroxyapatite (HAP) crystal growth(71). This

would be consistent with studies showing negative effects of NBPs on osteoblast matrix

mineralization. However, the structure of DGBP lacks the traditional R1 hydroxyl group

of the NBP structure. Clodronate, a non-NBP which similarly lacks the hydroxyl group,

exhibits lower inhibitory effects on HAP crystal growth in vitro than NBPs.

Concentrations less than or equal to 1 µM clodronate fail to inhibit osteoblast

mineralization(53, 54). This suggests that at the concentration tested in this study, DGBP

26

did not inhibit mineralization through a pyrophosphate-like effect on HAP crystal

growth; however, the binding affinity of DGBP for HAP has not been published.

As shown in Figure 4, exogenous GGPP inhibited osteoblast mineralization in the

presence or absence of DGBP suggesting the effect of DGBP to inhibit matrix

mineralization is not through GGPP depletion. This is consistent with data from Yoshida,

et al. showing that GGOH inhibits expression of osteoblastic genes and mineralization in

MC3T3-E1 pre-osteoblasts(55). We demonstrated that intracellular GGPP levels

decreased during MC3T3-E1 pre-osteoblast differentiation, in agreement with data from

Yoshida, et al. demonstrating decreased GGPPS expression and activity during MC3T3-

E1 pre-osteoblast differentiation(55). Interestingly, the GGPPS substrate FPP similarly

decreases during osteoblast differentiation, suggesting that isoprenoids upstream of

GGPP play a role in regulating osteoblast differentiation. This is consistent with work

published by Takase, et al. demonstrating decreased mevalonate kinase expression

following treatment of osteoblasts with the anabolic peptide PTH (56).

There have been many roles reported for FPP, in addition to being a precursor for

cholesterol synthesis and a substrate for protein prenylation. FPP has been reported to be

an activator of TRPv3(31) channels, an antagonist of the LPA3 receptor(30), and an

agonist of GPR92 orphan receptor(29). A study by Das, et al. showed that FPP activates

several nuclear hormone receptors, including the thyroid receptor, estrogen receptor, and

GR(27). Additionally, FPP activated the GR in epithelial cells and was found to play a

role in the regulation of wound healing(28). Figure 6 demonstrates that DGBP increased

the phosphorylation of the GR at serine residue 211, similar to the GR agonist

dexamethasone, and high concentrations of DGBP caused an increase in nuclear GR. The

27

decrease in cytosolic GR localization was evident with the GR agonist dexamethasone,

but not with DGBP. This may be due to the fact that dexamethasone is itself an agonist of

the GR, whereas activation of the GR by DGBP is likely indirect and dependent upon the

inhibition of GGPPS and the ensuing FPP accumulation. High doses of glucocorticoids

are known to have negative effects on the skeleton through direct effects on osteoblasts to

inhibit proliferation and differentiation(72). This suggests that activation of the GR by

FPP may play a role in the negative effects of DGBP on osteoblasts. Vukelic, et al. also

demonstrated that treatment with statins promotes wound healing through the depletion

of endogenous FPP(28). This calls to question whether statin drugs would similarly

decrease basal GR activity in pre-osteoblasts, which has been shown to be required for

osteoblast differentiation in murine models(72, 73). However, the low levels of activated

GR in the absence of dexamethasone or DGBP argues against a role for the low basal

levels of FPP in supporting GR activity. Taken together these results suggest that the

accumulation of FPP by GGPPS inhibition inhibits osteoblast differentiation potentially

through the activation of the GR, or other nuclear hormone receptors.

As mentioned previously, several publications have reported that a non-coding

SNP in the FPPS gene (rs2297480) correlates with BMD. Caucasian women with CC or

CA genotypes at this position displayed lower BMD at all sites measured as compared to

the AA genotype(61). Consistent with this, Marini, et al. 2008 demonstrated that post-

menopausal osteoporosis patients with the CC genotype showed a decreased response to

two years of NBP therapy as compared to the AA or AC genotypes (62). While the

functional consequence of this SNP has not been determined, this work highlights the

importance of FPP in the skeleton. It is possible that this SNP leads to higher expression

28

of FPPS and increased levels of FPP, potentially negatively affecting osteoblast

differentiation.

In summary, direct inhibition of GGPPS with DGBP depletes GGPP in

osteoblasts resulting in impaired geranylgeranylation. DGBP inhibits osteoblast matrix

mineralization and this is not prevented by GGPP add-backs, suggesting that the effect of

DGBP on matrix mineralization is independent of GGPP depletion. Consistent with the

effect of DGBP on matrix mineralization, DGBP inhibits expression of osteoblastic genes

in primary rat calvarial osteoblasts. Interestingly, inhibition of GGPPS led to an

accumulation of the GGPPS substrate FPP, and this accumulation remained over the

course of osteoblastic differentiation. This increase in FPP correlated with increased

phosphorylation and nuclear accumulation of the glucocorticoid receptor. Together these

results suggest a potential role for the depletion of isoprenoids upstream of GGPP, such

as FPP, in osteoblast differentiation.

29

Gene Entrez Gene

ID F/R Primer Sequence 5'-3'

Amplicon length (bps)

F AGACGCAGGTGTTCTTGGTCCTAA Adiponectin 246253

R 244

GAATTTGCCAGTGCTGCCGTCATA F AATCGGAACAACCTGACTGACCCT

ALP 25586 R

111 AATCCTGCCTCCTTCCACTAGCAA

F TGATCACCTGAACTCCACCAACCA BMP2 29373

R 176

AACCCTCCACAACCATGTCCTGAT F AGGCCAAGAAGTCGGTGGATAAGA

C/EPBα 24252 R

144 TGTCACTGGTCAACTCCAACACCT

F AGCAAAGGCAATGCTGAATCGTCC Col1a1 29393

R 125

TGCCAGATGGTTAGGCTCCTTCAA F TCCAGCACTTCTCCCAGATTGTCA FPPS 83791 R

193 AGGCTCTCAGCATCCTGTTTCCTT

F TGACTCTACCCACGGCAAGTTCAA GAPDH 24383

R 141

ACGACATACTCAGCACCAGCATCA F ATTTGGTCAAGGCCAGAAAGCACC GGPPS 291211 R

200 AAAGCCACTAGTGAAGGGTTCCCA

F TTCACGTTGGTCTCGGTGCTCTTA Glut4 25139

R 172

CCACAAAGCCAAATATGGCCACGA F TTCCAAGGGTACGGAGAAAGCACT HMGCR 25675 R

178 TTCTCTCACCACCTTGGCTGGAAT

F AGAACAGACAAGTCCCACACAGCA OCN 25295

R 185

TATTCACCACCTTACTGCCCTCCT F TCTCCAGCATTTCTGCTCCACACT PPARγ 25664 R

182 AGGCTCTTCATGTGGCCTGTTGTA

F TAAGGGACTCGAGGAGGTCAAGAA PPARγ2 25664

R 83

GGGAGTTAAGATGAATTTAGCGCTGC F AAGCACAAGTGATTGGCCGAACTG Runx2 367218 R

88 CCTCAACCACGAAGCCTGCAATTT

F ATTTCACAAGAATCAGGGCGTGGG Twist1 85489

R 111

ATCAGAATGCAGAGGTGTGAGGGT F ACTGGACCAAGGCTCTCAGAACAA

Twist2 59327 R

80 TTCCAGGCTTCCTCGAAACAGTCA

F AAAGCCATCTGCACGGGATCAAAC UCP1 24860

R 199

TCTGCCAGTATGTGGTGGTTCACA Table 1. Rat primers for real-time qPCR analysis.

30

Figure 3. DGBP reduces intracellular GGPP and impairs protein geranylgeranylation. A, Cells were treated with DGBP for 24 h. GGPP levels were quantified and normalized to total protein content. Data are expressed as pmoles per mg protein (% control). ap<.05 as compared to control MC3T3-E1 cells. bp<.05 as compared to control primary osteoblasts. B, MC3T3-E1 cells were treated with lovastatin or DGBP for 24 h. Total protein was isolated and used for Western blot analysis of protein prenylation. The Rap1a antibody probes for non-geranylgeranylated Rap1a, and a visible band indicates impaired geranylgeranylation. The Ras antibody probes for total Ras. The lower band is farnesylated Ras, and the upper band is non-farnesylated Ras. The existence of the upper band indicates impaired farnesylation of Ras. α-tubulin was used as a loading control.

31

Figure 4. DGBP inhibits osteoblast differentiation and matrix mineralization. A, MC3T3-E1 pre-osteoblasts were treated with 1µM DGBP in differentiation medium for 28 days. Alizarin red was used to detect mineralization. B, Mineralization was quantified by elution of Alizarin red and measurement of absorbance at 405nm. Data are expressed as µg Alizarin red per well (% control). ap<0.05 as compared to control, bp<0.05 as compared to DGBP treated osteoblasts. C, Cells were treated with 1µM DGBP for eight days. cDNA was transcribed from extracted mRNA and used for real time qPCR analysis of osteoblastic gene expression. Expression was normalized to the housekeeping gene GAPDH. *p<0.05 as compared to control.

32

Figure 5. DGBP leads to an accumulation of intracellular FPP. A, Cells were treated with indicated concentrations of DGBP. ap<0.05 as compared to control MC3T3-E1 cells. bp<0.05 as compared to control primary osteoblasts. B&C, MC3T3-E1 cells were treated with 1 µM DGBP for 1, 2, or 4 weeks. At each time point, intracellular levels of GGPP (B) and FPP (C) were assessed. ap<.05 as compared to week 1 controls, bp<.05 as compared to week 2 controls, cp<.05 as compared to week 4 controls. FPP levels were quantified and normalized to total protein content. Data are expressed as pmoles per mg protein (% control).

33

Figure 6. DGBP leads to activation of the glucocorticoid receptor. MC3T3-E1 cells were treated with 1.0 µM dexamethasone or DGBP (10, 50, and 100 µM) for 24 h. A, Whole cell lysate was extracted and Western blots were run to probe for total GR, P-GR (serine 211), and the housekeeping gene ß-tubulin. B, P-GR was normalized to total GR using densitometry. C, Nuclear/cytosol fractions were obtained and probed for total GR, Sp1 (nuclear), and ß-tubulin (cytosol). D, Levels of nuclear and cytosolic GR were quantified by densitometry and compared to control nuclear and cytosolic GR (% control).

34

CHAPTER III: EFFECTS OF FARNESYL PYROPHOSPHATE ACCUMULATION ON CALVARIAL OSTEOBLAST DIFFERENTIATION

Abstract

Statins, drugs commonly used to lower serum cholesterol, have been shown to

stimulate osteoblast differentiation and bone formation. Statins inhibit HMG-CoA

reductase (HMGCR), the first step of the isoprenoid biosynthetic pathway, leading to the

depletion of the isoprenoids farnesyl pyrophosphate (FPP) and geranylgeranyl

pyrophosphate (GGPP). The effects of statins on bone have previously been attributed to

the depletion of GGPP, since exogenous GGPP prevents statin-stimulated osteoblast

differentiation in vitro. However, in a recent report we demonstrated that the specific

depletion of GGPP did not stimulate, but in fact, inhibited osteoblast differentiation. This

draws the question of whether there are potential roles for other isoprenoids in the

regulation of osteoblast differentiation. We demonstrate that the expression of HMGCR

and FPP synthase (FPPS) decreased during primary calvarial osteoblast differentiation.

This correlated with decreased FPP and GGPP levels during differentiation. Zaragozic

acid (ZGA) inhibits the isoprenoid biosynthetic pathway enzyme squalene synthase

(SQS), leading to an accumulation of the SQS substrate FPP. ZGA treatment of calvarial

osteoblasts resulted in the inhibition of osteoblast differentiation as measured by

osteoblastic gene expression, alkaline phosphatase (ALP) activity, and matrix

mineralization. Osteoblast differentiation was restored by simultaneous HMGCR

inhibition, which prevented the accumulation of FPP. In contrast, specifically inhibiting

GGPP synthase (GGPPS) to lower the ZGA-induced increase in GGPP did not restore

osteoblast differentiation. Specificity of HMGCR inhibition to restore osteoblast

35

mineralization in ZGA treated cultures through the reduction in isoprenoid accumulation

was confirmed with mevalonate add-back experiments. Similar to ZGA treatment,

exogenous FPP inhibited the mineralization of primary calvarial osteoblasts.

Interestingly, the effects of FPP accumulation on osteoblasts were found to be

independent of protein farnesylation. Our findings are the first to demonstrate that the

accumulation of FPP impairs osteoblast differentiation and suggests that the depletion of

this isoprenoid may be necessary for normal and statin-induced bone formation.

Introduction

Statins, drugs commonly used to lower serum cholesterol, have several pleiotropic

effects. Mundy, et al. showed that these drugs could stimulate BMP-2 expression and

stimulate bone formation by osteoblasts(36). These results have been subsequently

supported by several in vitro(37-44) and in vivo(45, 46) studies. Additionally,

administration of lipophilic statins has been positively correlated with higher BMDs in

these patients(47-50).

Statins inhibit HMGCR, the first enzyme in the isoprenoid biosynthetic

pathway(33, 74) (Figure 2). Through inhibition of HMGCR, statins lead to the depletion

of FPP and GGPP, resulting in impaired protein prenylation. Current thought is that

statin-stimulated osteoblast differentiation occurs through the depletion of GGPP and

diminution of protein geranylgeranylation, since addition of exogenous GGPP or GGOH

prevent the effects of statins on bone(40, 43).

We demonstrated in Chapter II that, in contrast to our hypothesis, specific

inhibition of GGPPS by DGBP, which depleted GGPP, led to inhibition of osteoblast

36

differentiation. We also noted that FPP levels decrease during MC3T3-E1 pre osteoblast

differentiation and this was prevented by DGBP treatment. In this study we sought to

determine whether isoprenoids upstream of GGPP, specifically FPP, have a negative

effect on osteoblast differentiation, potentially playing a role in the regulation of bone

formation.

Materials and Methods

Primary cell isolation and culture. The MC3T3-E1 pre-osteoblast cell line was

obtained from ATCC. The animal protocol used for isolation of primary cells was

approved by the Institutional Animal Care and Use Committee at the University of Iowa.

Primary rat osteoblast cells were obtained by three sequential enzyme digestions of

calvariae from two day old neonatal Sprague-Dawley rats (Harlan). Digestions were

performed with 0.05% collagenase type I (Sigma) and 1% trypsin (Invitrogen) in serum

free α modified essential medium (α-MEM) (Invitrogen) at 37oC with shaking. The first

two digestions (10 and 20 minutes, respectively) were discarded. The last digestion (60

minutes) was collected and cells were centrifuged and resuspended in α-MEM containing

10% fetal bovine serum (FBS) and 1x penicillin-streptomycin (Invitrogen). Cells were

grown in a humidified atmosphere with 5% CO2 at 37oC in 10 cm plates. Upon

confluency cells were sub-cultured for experiments at a density of 1x104 cells/cm2.

All experiments were carried out in osteoblast differentiation medium. This

consisted of -MEM with 10 mM β-glycerophosphate (Sigma) and 50 µg/mL L-ascorbic

acid (Fischer Scientific). Compounds utilized for experimental treatments included

zaragozic acid (Sigma), lovastatin (Sigma), mevalonate (Sigma), farnesyl transferase

37

inhibitor (FTI)-277 (Sigma), or digeranyl bisphosphonate (DGBP). SQSI-154 (or LWS-

154), an alternative squalene synthase inhibitor synthesized in Dr. David F. Wiemer’s lab

at the University of Iowa, was also tested with MC3T3-E1 pre-osteoblast cultures.

Treatments were replaced every 3-4 days until the termination of the experiment.

Zaragozic acid, FTI-277, SQSI-154, and DGBP were dissolved in water. Lovastatin and

mevalonic acid lactone were subjected to lactone hydolysis, followed by dilution with

water or RPMI medium, respectively.

Cell Proliferation Assay. Cells were plated in 24 well plates. Upon confluency,

cells were treated for 72 h. 3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide

(MTT) (Calbiochem) was added to cells and the reaction was incubated at 37oC with 5%

CO2 for 3 h. The reaction was terminated with MTT stop solution (HCl, tritonX-100, and

isopropyl alcohol). Plates were shaken overnight at 37oC. Absorbance was measured at

540 nm with a reference wavelength of 650 nm.

Real-time quantitative polymerase chain reaction (qPCR). Primary calvarial

osteoblasts were plated in 6-well plates. Treatment was applied when cells reached

confluency. Four to five days following the onset of treatment, total RNA was isolated

from each well using Qiashredders and the RNeasy Mini Kit (Qiagen). During the

isolation, a DNase step was performed (Qiagen). One µg of RNA was reverse-transcribed

into cDNA using the iScript cDNA Synthesis Kit (Bio-Rad). Real-time qPCR was

performed with Sybr Green Master Mix (Applied Biosystems) on an ABI SDS 7900 HT

(Applied Biosystems). The real-time protocol consisted of 2 minutes at 50oC, 10 minutes

at 95oC, followed by 40 cycles of 95oC (15 seconds) and 60oC (one minute).

Dissociation curves were obtained following the real-time qPCR to ensure the proper

38

amplification of target cDNAs. Primers were obtained from Integrated DNA

Technologies (Iowa City, IA, USA) and eluted in TE Buffer (Ambion). Sequences and

amplicon lengths are found in the Table 1.

Mineralization assay. Primary calvarial osteoblasts were plated in 24 well plates.

Treatment began when cells reached confluency. After 14 days of culture, cells were

fixed for one hour in ice cold 70% ethanol. Cells were then washed thoroughly with DI

H2O; mineralization was detected with 40 mM Alizarin red (Sigma), pH=4.2 for 15

minutes. Following staining, cells were washed thoroughly with DI H2O to remove non-

specific Alizarin red. Images were captured using a Canon EOS Rebel XS.

Mineralization was quantified as described previously(70). Briefly, Alizarin Red was

eluted with 10% acetic acid. Plates were shaken on an orbital rotator for 15 minutes.

The acetic acid and cells from each well were transferred to a 1.5 mL tube and heated to

85oC for 10 minutes. The samples were then cooled on ice and centrifuged. The

supernatant was transferred to a new tube and 10% ammonium hydroxide was added.

Samples were vortexed vigorously and aliquots were transferred to a 96 well plate. The

absorbance was read at 405 nm on a Thermomax Microplate Reader (Molecular

Devices). Absorbance was compared to an Alizarin red standard curve; values were

expressed as total mol Alizarin Red per well.

Alkaline Phosphatase (ALP) activity assay. Calvarial osteoblasts were plated in

12-well plates. One week following the onset of treatment, cells were washed twice with

PBS (Invitrogen). Cells were lysed with 0.2% Triton-X 100 (Sigma) and subjected to two

freeze thaw cycles. Cells were transferred to 1.5 mL tubes and centrifuged. The

supernatants were utilized in the ALP assay. Five mg ALP substrate tablets (pNPP)

39

(Sigma) were dissolved in alkaline buffer (Sigma) (40mg/10mL). Cell lysate or control

lysis buffer was transferred to a 96 well plate and substrate solution was added to each

well. The assay was carried out at 37oC for 10 minutes. Absorbance was read at 405 nm

on a Thermomax Microplate reader (Molecular Devices). ALP activity (units/mL) was

calculated from a standard curve created with 4-nitrophenol. ALP activity was

normalized to total protein content measured by BCA assay. Values are expressed as

ALP units per mg protein per minute.

FPP/GGPP Quantitation. Calvarial osteoblasts were plated in 10cm plates. Upon

confluency cells were treated for 24 hours. FPP and GGPP levels were determined as

previously reported by reverse phase HPLC(69). Briefly, cells were washed twice with

PBS (Invitrogen) and isoprenoid pyrophosphates were extracted from cells and used as

substrates for incorporation into fluorescent GCVLS or GCVLL peptides by

farnesyltransferase or geranylgeranyl transferase I. The prenylated fluorescent peptides

were separated by reverse phase HPLC and quantified by fluorescence detection. Total

FPP and GGPP levels were normalized to total protein content as measured by BCA

assay. Values are expressed as pmoles FPP or GGPP per mg protein.

Statistical Analysis. Data are expressed as means ± SEM. All experiments were

repeated at least twice with similar results. Differences between two groups were

compared using unpaired student’s t tests. Statistical significance was defined by p values

less than 0.05.

40

Results

The expression of isoprenoid biosynthetic pathway enzymes decreases during

osteoblast differentiation. To determine intracellular levels of the isoprenoids FPP and

GGPP during osteoblast differentiation, calvarial osteoblasts were treated with

differentiation medium for 0, 4, 7, or 10 days. At each end point intracellular isoprenoids

were extracted and quantified by HPLC as described in the methods. Both FPP and

GGPP decreased over the course of osteoblast differentiation (Figure 7A).

Expression of isoprenoid pathway enzymes was also assessed. Calvarial

osteoblasts were treated with differentiation medium for 0, 2, 4, or 7 days. At each end

point cells were analyzed for expression of the isoprenoid pathway enzymes HMGCR,

FPPS, and GGPPS. Expression of HMGCR and FPPS decreased significantly during

osteoblast differentiation (Figure 7B). In contrast, GGPPS levels increased significantly

at day 7. Differentiation was assessed by expression of the mature osteoblast marker

OCN (Figure 7C).

ZGA leads to an accumulation of FPP. To determine the effect of the squalene

synthase inhibitor, zaragozic acid (ZGA) (Figure 2) on accumulation of intracellular FPP

and GGPP levels in osteoblasts, calvarial osteoblasts were treated with ZGA (1-10 µM).

FPP and GGPP were both increased following treatment with ZGA (Fig. 8). GGPP

increased maximally with one M ZGA. In contrast, FPP, which was increased

significantly with one µM ZGA, increased greater with higher doses of ZGA (5-10 M).

Control levels of FPP and GGPP in primary calvarial osteoblasts were 3.8 ± 0.3 and 1.5 ±

0.3 pmol per mg protein, respectively.

41

ZGA inhibits osteoblast mineralization and osteoblast gene expression. To

examine whether ZGA inhibits matrix mineralization, calvarial osteoblasts were treated

with 1-10 µM ZGA for two weeks followed by Alizarin red staining of calcium

deposition. As shown in Figure 9A and B, ZGA significantly inhibited matrix

mineralization, with 5 and 10 µM ZGA resulting in nearly total inhibition of matrix

mineralization.

The effect of ZGA on osteoblast viability was assessed by MTT assay. Cells were

treated with 0.5-25 µM ZGA for 72 hours. As shown in Figure 9C, maximal inhibition of

proliferation was approximately 40% which occurred with 5-25 µM ZGA.

To determine whether ZGA inhibits osteoblast differentiation, ALP activity and

expression of osteoblast differentiation markers were analyzed. For analysis of

expression, cells were treated for four days with 1-10 µM ZGA. Figure 9D shows the

inhibition of expression of osteoblast markers ALP, BMP-2, and OCN following four

days of treatment. Transcription factor Runx2 and the late osteoblastic gene Col1a1 were

affected to a lesser extent. For analysis of ALP activity, cells were treated with 1-10 µM

ZGA for seven days. Consistent with decreased expression of ALP, ZGA inhibited ALP

activity in a concentration-dependent manner (Figure 9E).

SQSI-154 causes an accumulation of FPP and inhibits MC3T3-E1 pre-osteoblast

differentiation. To confirm the effects of FPP accumulation on osteoblasts we utilized a

novel SQS inhibitor, SQSI-154 synthesized by Dr. David F. Wiemer’s lab at the

University of Iowa. Treatment of MC3T3-E1 pre-osteoblasts with SQSI-154, otherwise

referred to as LWS-154, led to a concentration-dependent accumulation of FPP (Figure

10A), similar to ZGA treatment of primary rat calvarial osteoblasts. MC3T3-E1 pre-

42

osteoblasts were cultured for 28 days in the presence or absence of 0.5-5.0 M SQSI-154.

SQSI-154 inhibited osteoblast matrix mineralization at all concentrations tested as

measured by the elution of Alizarin red (Figure 10B).

Inhibition of osteoblast differentiation and mineralization by ZGA is prevented by

inhibition of HMGCR but not GGPPS. To assess whether the effect of ZGA to inhibit

osteoblast differentiation and matrix mineralization is due specifically to the

accumulation of isoprenoids, lovastatin co-treatments were utilized. As described earlier,

statins inhibit HMGCR resulting in the depletion of mevalonate and its subsequent

metabolites, including FPP and GGPP. Figure 11A demonstrates that 1.0 µM lovastatin

alone led to a reduction in intracellular FPP and GGPP. Five micromolars ZGA, as shown

earlier, led to a significant increase in intracellular FPP levels. Co-treatment of 1 µM

lovastatin with 5 µM ZGA led to a significant increase in FPP as compared to control;

however, the FPP accumulation was significantly decreased as compared to ZGA alone.

ZGA also increases GGPP levels (Fig 8). At a concentration of 1.0 µM digeranyl

bisphosphonate (DGBP), which specifically targets GGPPS (64, 65), the enzyme

downstream of FPPS (Figure 2), led to a slight accumulation of FPP. Co-treatment of 5.0

µM ZGA and 1.0 µM DGBP reduced the accumulation of GGPP (Figure 11B). As

expected, there was no significant reduction in FPP accumulation as compared to ZGA

alone (Figure 11A).

Figures 11C-E show that the lovastatin concentration chosen for ZGA co-

treatments (1.0 µM) alone had only minimal effects on matrix mineralization and

increased ALP activity of calvarial osteoblasts, with no effect on osteoblast viability.

Similar to previous experiments, 5.0 µM ZGA inhibited osteoblast viability, ALP

43

activity, and mineralization. Co-treatment of 5.0 µM ZGA with 1.0 µM lovastatin

resulted in a significant restoration of matrix mineralization, osteoblast viability, and

ALP activity as compared to ZGA alone. This suggests that the effect of ZGA to inhibit

osteoblast differentiation is due to the accumulation of isoprenoids.

To confirm that the inhibition of osteoblast differentiation and matrix

mineralization by ZGA was not due to the increase in GGPP, DGBP co-treatments were

utilized. Alone, 1.0 µM DGBP led to a decrease in ALP activity (Figure 11E) and matrix

mineralization (Figure 11C), with no effect on osteoblast viability (Figure 11D). In

contrast to lovastatin, DGBP did not prevent the inhibition of matrix mineralization,

osteoblast viability, and ALP activity by ZGA treatment of primary calvarial osteoblasts

(Figures 11C-E). Higher concentrations of DGBP (5.0 µM) were toxic in the absence of

ZGA. In the presence of 5.0 M ZGA, 5 M DGBP led to a reduction of GGPP below

control levels (Figure 12A); however, similarly to the 1.0 µM DGBP co-treatment, these

concentrations did not restore osteoblast matrix mineralization (Figure 12B). Treatment

of primary calvarial osteoblasts with 5-20 µM exogenous FPP for two weeks led to

decreased matrix mineralization (Figure 12C).

Restoration of osteoblast differentiation by lovastatin co-treatment is due to

prevention of isoprenoid accumulation. To determine whether the restoration of

osteoblastic differentiation and matrix mineralization by lovastatin co-treatments was

through the prevention of isoprenoid accumulation, mevalonate (Mev) add-back

experiments were performed. Addition of 5 mM Mev alone had no effect on matrix

mineralization as measured by Alizarin red staining (Figure 13A). ZGA inhibited matrix

mineralization and this was restored by co-treatment with 5.0 µM lovastatin. Addition of

44

5.0 mM Mev to the ZGA and lovastatin co-treatment significantly decreased

mineralization as compared to ZGA or lovastatin alone. This data suggests that the

restoration of osteoblast differentiation by lovastatin is due to the prevention of

isoprenoid accumulation and not an off target effect of lovastatin.

The effect of FPP accumulation on inhibition of osteoblast differentiation is not

due to increased protein farnesylation. To test whether the negative effects of FPP

accumulation on osteoblast differentiation are due to increased protein farnesylation,

farnesyltransferase inhibitor (FTI) co-treatments were done. FTI-277 was employed at

concentrations of 0.1, 1.0, and 10.0 µM alone or in combination with 5 µM ZGA. As

shown in Fig. 13B, FTI-277 had no significant effect on osteoblast mineralization at any

concentration tested. In combination with ZGA, FTI-277 did not have any significant

effect on ZGA-induced inhibition of osteoblast differentiation. This suggests that the

accumulation of isoprenoids, such as FPP, act independently of farnesylation to inhibit

osteoblast differentiation.

Discussion

The results presented herein demonstrate that isoprenoids upstream of GGPP,

specifically FPP, have negative effects on osteoblast differentiation. This was evidenced

by ZGA treatment, which increased FPP and GGPP isoprenoid levels, leading to

inhibition of osteoblast differentiation, which could be prevented by inhibition of

HMGCR, but not that of GGPPS.

Together with the results presented in Chapter II, these data have important

implications for the numerous studies reporting positive effects of statins on osteoblast

45

differentiation. As described earlier, several in vitro(37-44) and in vivo(45, 46)

experiments have shown positive effects of statins on osteoblast differentiation and bone

formation. These findings have been attributed to the depletion of GGPP (40, 43).

Consistent with this paradigm, several studies have found that nitrogenous

bisphosphonates (NBPs), which inhibit FPPS, (13, 15, 75), similarly leading to GGPP

depletion (Figure 2), stimulate osteoblast differentiation (51, 52, 76-80). It is important to

note however that the effects of NBPs on osteoblasts are unclear, as some studies have

shown the ability to inhibit matrix mineralization and cause osteoblast apoptosis(53, 54,

81, 82).

Our studies show that the isoprenoids FPP and GGPP, as well as the expression of

HMGCR and FPPS, decrease during primary calvarial osteoblast differentiation. This is

similar to the decrease in FPP and GGPP levels demonstrated during the differentiation

of MC3T3-E1 pre-osteoblasts, described in Chapter II. However, in contrast to the

previous study by Yoshida, et al(83), expression of GGPPS increased with primary

osteoblast differentiation. This increased expression may be occurring in order to

compensate for the decreased intracellular GGPP levels during osteoblast differentiation.

Unlike HMGCR and FPPS, GGPPS expression is not regulated by SREBP2, suggesting

that GGPPS expression may be modulated by a separate signal, such as intracellular

GGPP levels. The results showing decreased expression of HMGCR and FPPS, as well as

decreased FPP during osteoblast differentiation, suggest that depletion of more than just

GGPP plays a role in osteoblast differentiation. Consistent with this, Takase, et al. show

that the anabolic protein PTH leads to a reduction in mevalonate kinase expression(56). It

46

is important to note that the results presented in this study and those of Yoshida, et al. are

in agreement that GGPP levels decrease during osteoblast differentiation (83).

ZGA treatment led to an increase in intracellular FPP and GGPP and decreased

osteoblast differentiation and matrix mineralization. We concluded that the negative

effects of ZGA on osteoblast differentiation were due to accumulation of upstream

isoprenoid metabolites, specifically FPP, not GGPP. Several lines of evidence support

this conclusion. GGPP increased maximally at a concentration of 1.0 µM ZGA; however

inhibition of osteoblast gene expression and ALP activity did not occur with this

concentration. While 1.0 µM ZGA did decrease matrix mineralization, maximal

inhibition of matrix mineralization, did not occur with this concentration. Secondly,

addition of DGBP, which prevented the accumulation of GGPP in ZGA-treated cultures,

did not prevent the inhibition of osteoblast viability, ALP activity, or matrix

mineralization. In contrast, addition of lovastatin, which inhibits HMGCR thereby

preventing the accumulation of FPP (and likely other upstream isoprenoids which were

not measured in this study), restored osteoblast viability, ALP activity, and matrix

mineralization. These experiments suggest that ZGA inhibits osteoblast differentiation

and matrix mineralization through the accumulation of upstream isoprenoid metabolites.

Additionally, treatment with exogenous FPP inhibited matrix mineralization, suggesting

that the accumulation of FPP is responsible for the negative effects of ZGA. The effects

of FPP on matrix mineralization were less than those of ZGA. This is likely due to this

isoprenoid’s poor cellular permeability, which limits the intracellular concentrations of

FPP that accumulate with exogenous FPP treatment.

47

There has been much study recently into potential roles for FPP aside from

protein farnesylation and the production of sterols and non-sterol isoprenoids. Das, et al.

showed that FPP binds and activates several nuclear receptors, including the thyroid

hormone receptor, estrogen receptor, and glucocorticoid receptor(27). In a more recent

report, Vukelic, et al. demonstrated that modulation of FPP levels altered wound healing

through modulation of glucocorticoid receptor activities(28). We demonstrated in

Chapter II that DGBP, which also causes an accumulation of FPP (Figure 5A), activates

the glucocorticoid receptor in MC3T3-E1 pre-osteoblasts (Figure 6). The potential role of

FPP and other isoprenoid metabolites to activate nuclear hormone receptors remains a

potential mechanism of ZGA-mediated inhibition of osteoblast differentiation and matrix

mineralization. In support of a non-prenylation role for FPP in inhibiting osteoblast

differentiation, the effects of ZGA on osteoblast mineralization were not prevented by

inhibition of farnesylation with FTI-277.

In summary, we demonstrate for the first time in this report that the expression of

isoprenoid pathway enzymes HMGCR and FPPS are down-regulated during osteoblast

differentiation, and the accumulation of isoprenoids upstream of GGPP, including FPP,

impairs osteoblast differentiation. Our results suggest a role for the depletion of FPP in

normal and statin-stimulated osteoblast differentiation.

48

Figure 7. The isoprenoid biosynthetic pathway and primary calvarial osteoblast differentiation. A, Intracellular FPP and GGPP decrease during osteoblast differentiation. Data are expressed as pmoles per mg protein, percent day 0 (mean ± SEM), n=2. B, Expression of the isoprenoid pathway enzymes HMGCR and FPPS decrease during osteoblast differentiation. mRNA levels were quantified by real-time qPCR and were normalized to the reference gene GAPDH. Data are expressed as relative units (mean ± SEM). *p<.05 versus undifferentiated cells (day 0), n=3. C, OCN expression increased on days 2-7 showing the differentiation of the calvarial osteoblasts. Expression was normalized to the reference gene GAPDH, and data was expressed as relative units (mean ± SEM), n=3.

49

Figure 8. ZGA leads to increases in endogenous FPP and GGPP. A, Primary calvarial osteoblasts were treated with increasing doses of ZGA (1-10 µM). FPP and GGPP are expressed as pmoles per mg protein fold vehicle (mean ± SEM). ap<0.05 versus control, bp<0.05 versus 1 µM ZGA, n=2.

50

Figure 9. ZGA leads to inhibition of osteoblast differentiation. A,B ZGA concentration-dependently (1-10 µM) inhibited matrix mineralization of calvarial osteoblasts as assessed by Alizarin red staining (A) and quantification of eluted dye (B). Data are expressed as micromoles Alizarin red per well, percent vehicle (mean ± SEM). *p<0.05 versus control treated cells, n=3. C, ZGA inhibited cell viability as measured by MTT assay. Values are expressed as percent control (mean ± SEM), n=3. D, ZGA inhibited expression of osteoblastic differentiation markers ALP, BMP2, and OCN as measured by real-time qPCR. Partial inhibition of expression was observed for Col1a1 and Runx2. Data are normalized to the reference gene GAPDH and expressed as percent vehicle (mean ± SEM). *p<0.05 versus control treated cells, n=3. E, ZGA inhibited ALP activity of calvarial osteoblasts. ALP activity was assayed and normalized to protein content. Data are reported as ALP units/mg protein/minute, percent vehicle (mean ± SEM). *p<0.05 as compared to control treated cells, n=3.

51

52

Figure 10. SQSI-154 causes an accumulation of FPP and inhibits osteoblast matrix mineralization. A, MC3T3-E1 pre-osteoblasts were treated with increasing concentrations of SQSI-154 (0.5-5 µM). FPP and GGPP are expressed as pmoles per mg protein, fold vehicle (mean ± SEM), n=2. B, SQSI-154 inhibited matrix mineralization of MC3T3-E1 pre-osteoblasts as assessed by quantification of Alizarin red staining. Data are expressed as moles Alizarin red per well, percent vehicle (mean ± SEM), n=3

53

Figure 11. Inhibition of HMGCR prevents FPP accumulation and inhibition of osteoblast differentiation in cells treated with ZGA. Primary calvarial osteoblasts were treated with 5µM ZGA in the presence or absence of an inhibitor to HMGCR (Lov 1 µM) or GGPPS (DGBP 1 µM). A, The accumulation of FPP by ZGA was prevented by inhibition of HMGCR, but not inhibition of GGPPS. Values in A and B are expressed as pmoles per mg protein, fold vehicle (mean ± SEM). ap<0.05 as compared to control, bp<0.05 as compared to ZGA treated cells, n=2. B, 1 M DGBP reduced the accumulation of GGPP (mean SEM), n=2 C, Inhibition of HMGCR prevented the inhibition of mineralization by ZGA, whereas inhibition of GGPPS did not. Osteoblast mineralization was assessed by Alizarin red staining followed by elution and quantification of dye. Values were expressed as µg Alizarin red per well (mean ± SEM). ap<0.05 as compared to control, bp<0.05 as compared to ZGA treated cells, n=3. D, Cell viability was assessed by MTT assay and expressed as percent vehicle (mean ± SEM), n=3. Additionally, E, ALP activity was assessed. Values were normalized to protein content. Data were expressed as percent vehicle, ALP units per mg protein per minute (mean ±S EM). ap<0.05 as compared to vehicle, bp<0.05 as compared to ZGA treated cells, n=3.

54

55

Figure 12. GGPP depletion does not restore osteoblast mineralization. A and B, Primary calvarial osteoblasts were treated with 5 µM ZGA and/or 5 µM DGBP. A, GGPP levels were quantified following 24 hours of treatment and are expressed as pmole per mg protein, percent vehicle. ap<0.05 as compared to vehicle control, bp<0.05 as compared to 5 µM ZGA alone, (mean ± SEM) n=2. B, Cells were treated for 14 days and mineralization was assessed by Alizarin red staining and elution of the stain. Data are expressed as percent vehicle (mean ± SEM), *p<0.05 as compared to vehicle, n=3. C, Primary calvarial osteoblasts were treated with FPP for two weeks. Mineralization was assessed by Alizarin red staining.

56

Figure 13. Addition of exogenous mevalonate and co-treatment with FTI-277. A, Cells were treated with ZGA, Lov, or ZGA/Lov combination in the presence or absence of 5 mM exogenous mevalonate (Mev). Osteoblast mineralization was assessed by Alizarin red staining followed by elution and quantification of dye. Values were expressed as µg Alizarin red per well (mean ± SEM). ap<0.05 as compared to control, bp<.05 as compared to ZGA treated cells, cp<.05 as compared to ZGA/Lov control, n=3. B, Cells were treated with 5 µM ZGA in the presence or absence of FTI-277 (0.1-10 µM). Osteoblast mineralization was assessed by Alizarin Red staining followed by elution and quantification of dye. Values were expressed as µg Alizarin red per well, percent control (mean ± SEM), n=3.

57

58

CHAPTER IV: GGPP AND THE OSTEOBLAST VS. ADIPOCYTE FATE DECISION Abstract Osteoblasts and adipocytes are derived from mesenchymal stem cells and play

important roles in the homeostasis of the skeleton. During the osteoblastic differentiation

of MC3T3-E1 pre-osteoblasts and primary calvarial osteoblasts there is a decrease in the

isoprenoid geranylgeranyl pyrophosphate (GGPP). Consistent with this, osteoblast

differentiation has been shown to be stimulated by statin drugs through the depletion of

GGPP. In contrast, adipogenic differentiation of 3T3-L1 cells results in increased

expression of GGPP synthase (GGPPS), and GGPP lowering agents inhibit adipogenesis

in vitro. In this study, we tested the hypothesis that GGPP levels play a role in the

osteoblast versus adipocyte fate decision by inhibiting osteoblast differentiation and

enhancing adipogenesis. We found that exogenous GGPP reduced osteoblastic gene

expression and matrix mineralization in primary calvarial osteoblast cultures. GGPP

treatment of primary calvarial osteoblasts and primary bone marrow stromal cells

(BMSCs) also led to the increased expression of total peroxisome proliferator activated

receptor (PPAR)-γ as well as the PPARγ2 splice variant. Inhibition of PPARγ

transcriptional activity did not prevent the effects of GGPP on osteoblast differentiation.

Enhanced PPARγ expression correlated with the increased formation of Oil Red O-

positive cells in these cultures. Additionally, primary calvarial osteoblasts treated with

GGPP exhibited increased expression of the adipokine adiponectin. In contrast to

previous reports utilizing other cell types, treatment of osteoblasts with GGPP did not

increase geranylgeranylation, suggesting that GGPP itself may be acting as a signaling

molecule. GGPP treatment of MC3T3-E1 pre-osteoblasts and primary calvarial

59

osteoblasts led to enhanced insulin-induced Erk signaling and increased glucose uptake;

it is important to note, however, that there was a decrease in the phosphorylation of the

insulin receptor. Altogether these findings demonstrate a negative role for GGPP in

osteoblast differentiation, leading to increased adipogenesis. Additionally, the effects of

GGPP on insulin signaling and glucose uptake implicate a role for this isoprenoid in

physiological energy homeostasis.

Introduction As discussed in previous chapters, statins have been shown to stimulate osteoblast

differentiation and bone formation(36-44). Statins act by inhibiting HMGCR, the first

step of the isoprenoid biosynthetic pathway and the rate-limiting step of cholesterol

biosynthesis (33). By inhibiting HMGCR, statins also deplete GGPP, leading to impaired

protein geranylgeranylation. It has been thought that the positive effects of statins on

osteoblasts are due to the depletion of GGPP, since addition of this isoprenoid prevents

these effects in vitro. While we show in Chapter II that specific depletion of GGPP does

not lead to osteoblast differentiation, there is evidence that GGPP does have a negative

role in this process. We have demonstrated in Chapters II and III that GGPP levels

decrease during osteoblast differentiation in the cases of both primary calvarial

osteoblasts(84) and MC3T3-E1 pre-osteoblasts(85). We also demonstrated that GGPP

addition to MC3T3-E1 cell cultures inhibited matrix mineralization(85) consistent with

another study showing that GGOH, which can be converted to GGPP by intracellular

kinases, inhibits osteoblast differentiation in MC3T3-E1 pre-osteoblast cultures(55).

Osteoblasts are derived from mesenchymal stem cells (MSCs). In studying the

differentiation of osteoblasts and bone formation, much attention is paid to the

60

differentiation of another mesenchymal-derived cell type involved in bone homeostasis,

the adipocyte (2, 57). Although initially thought to be inert(6, 58, 86), adipocytes are now

known to be central players in the control of energy balance and whole body lipid

homeostasis. Through their secreted adipokines, adipocytes play a role in obesity,

atherogenesis, diabetes, and inflammation(57, 58). Differentiation into osteoblasts or

adipocytes is controlled by the expression of lineage specific transcription factors, such

as Runx2 or PPARγ2, respectively(57). Interestingly, factors that induce adipogenesis

inhibit osteoblast differentiation and vice versa. For example, PPAR2 ligands, such as

the thiazolidinedione anti-diabetic drugs, promote adipogenesis and inhibit osteoblast

differentiation(2, 57, 58, 87, 88).

In the identification of mammalian GGPPS, Vicent, et al. found that this

isoprenoid biosynthetic enzyme was highly expressed in the liver, skeletal muscle, and

adipose tissue of the ob/ob mice, a model of leptin deficiency. Additionally, Vicent, et al.

demonstrated that GGPPS expression increases during the adipogenic differentiation of

3T3-L1 cells (60). Consistent with this, statins, which as mentioned above deplete GGPP,

have been shown to inhibit adipogenesis(42, 59).

Given these data, we hypothesized that GGPP may influence the osteoblastic

versus adipogenic fate decision, promoting adipogenesis and inhibiting

osteoblastogenesis. Herein we demonstrate the negative effects of GGPP on osteoblasts

and propose potential mechanisms by which GGPP exerts these effects. We also

demonstrate for the first time an adipogenic enhancing effect of GGPP.

61

Materials and Methods

Cell culture. The MC3T3-E1 pre-osteoblast cell line was obtained from ATCC.

The animal protocol used for isolation of primary rat calvarial cells and primary rat bone

marrow stromal cells was approved by the Institutional Animal Care and Use Committee

at the University of Iowa. Primary rat osteoblasts were obtained by three sequential

enzyme digestions of calvariae from two day old neonatal Sprague-Dawley rats (Harlan).

Digestions were performed with 0.05% collagenase type I and 1% trypsin (Invitrogen) in

serum free α modified essential medium (α-MEM) (Invitrogen) at 37 oC with shaking.

The first two digestions (10 and 20 minutes, respectively) were discarded. Following the

final digestion (60 minutes), cells were centrifuged and resuspended in growth medium.

Primary rat bone marrow stromal cells (BMSCs) were obtained by isolation of the femur

and tibia from five-week-old male rats. The epiphyses were removed in sterile conditions

and the bone was flushed with α-MEM. The cell mixture was strained to remove debris.

The cell mixture was washed, centrifuged, and resuspended in growth medium. Cells

were cultured in α-MEM containing 10% fetal bovine serum (FBS) and 1x penicillin-

streptomycin (Invitrogen). Cells were grown in a humidified atmosphere with 5% CO2 at

37oC in 10 cm plates.

Cells were subcultured for experiments at a density of 1x104 cells/cm2. Primary

calvarial osteoblast experiments were carried out in osteoblast differentiation medium,

consisting of -MEM with 10 mM β-glycerophosphate (Sigma) and 50 µg/mL L-

ascorbic acid (Fischer Scientific). Primary BMSC culture experiments were carried out in

control, osteoblast, or adipocyte differentiation medium. BMSC osteoblast differentiation

medium consisted of -MEM with 10 mM β-glycerophosphate (Sigma), 50 µg/mL L-

62

ascorbic acid (Fischer Scientific), and 0.1 µM dexamethasone (Sigma). Adipocyte

differentiation medium consisted of -MEM with 0.5 µg/mL insulin, 0.5 mM

isobutylmethylxanthine, 200 µM indomethacin, and 1.0 µM dexamethasone (all

purchased from Sigma). Treatments were replaced every 3-4 days until the termination of

the experiment.

FPP/GGPP Quantification. Primary BMSCs were plated in 10 cm plates. Upon

confluency cells were treated with DGBP (Dr. David F. Wiemer laboratory, University of

Iowa) (0.1, 1, 10, 100 µM) for 24 h. FPP and GGPP levels were determined as previously

reported by reverse phase HPLC(69). Briefly, cells were washed twice with phosphate

buffered saline (PBS) (Invitrogen) and isoprenoid pyrophosphates were extracted from

cells and used as substrates for incorporation into fluorescent GCVLS or GCVLL

peptides by farnesyltransferase or geranylgeranyl transferase I. The prenylated

fluorescent peptides were separated by reverse phase HPLC and quantified by

fluorescence detection. Total FPP and GGPP levels were normalized to total protein

content as measured by bichinconic (BCA) assay. Values are expressed as pmoles per mg

protein.

Western blotting. MC3T3-E1 pre-osteoblasts and primary calvarial osteoblasts

were plated in 10 cm plates. Upon confluency, cells were serum-starved overnight. The

cells were pre-treated with GGPP (20 µM) or GGPP vehicle (MeOH:10mM NH4OH, 7:3)

for 30 minutes, followed by treatment with 100 nM insulin for 0, 5, 10, 15, or 30 minutes.

At the end of the experiment, media was removed and cells were washed twice in ice-

cold PBS. Cells were collected with a cell scraper after the addition of RIPA buffer

supplemented with protease inhibitor cocktail, sodium vanadate, sodium fluoride, and

63

phenylmethylsulphonyl fluoride. Lysates were transferred to a 1.5 ml tube, incubated on

ice for 30 min, and passed through a 27-guage needle. Lysates were then centrifuged and

supernatant was transferred to a fresh 1.5 ml tube. Protein concentrations were

determined by the BCA method. Proteins were resolved on 7.5% SDS-PAGE gels and

transferred to polyvinylidene difluoride membranes by electrophoresis. Membranes were

blocked in 0.5% BSA (insulin receptor, phosph-insulin receptor, insulin receptor

substrate (IRS), phospho-IRS, Akt, and phospho-Akt antibodies) or 0.5% skim milk (α-

tubulin, phospho-Erk, Erk). Primary antibody was added to the blocking mixture and

membranes were rotated at 4oC overnight. Membranes were washed and secondary

antibodies were added in 0.5% skim milk for 1 h at 37oC. Proteins were visualized using

an enhanced chemiluminescence detection kit from GE Healthcare. Insulin receptor

(InsR)-, phospho-InsR- (Tyr1150,1151), IRS-1, phospho-IRS-1(Ser612), Akt, and

phospho-Akt antibodies were obtained from Cell Signaling. Antibodies to α-tubulin,

Erk1/2, phospho-Erk1/2, and Rab6 were obtained from Santa Cruz Biotechnology, Inc.

Horseradish peroxidase-conjugated anti-mouse and anti-rabbit secondary antibodies were

purchased from GE Healthcare.

Triton X-114 separation. MC3T3-E1 pre-osteoblasts were plated in 10 cm plates.

When the cells reached confluency they were treated for 24 h with 20 µM lovastatin,

GGPP vehicle, 10 µM GGPP, or 20 µM GGPP. Cells were washed twice with ice-cold

PBS and lysed in ice-cold Triton X-114 lysis buffer (20 mM Tris pH 7.5, 150 mM NaCl,

and 1% Triton X-114). Cell lysate was then passed through a 27-gauge needle,

centrifuged, and transferred to a new tube as described above. Separation of aqueous and

detergent phases was performed as described previously(89). Briefly, cells were

64

incubated at 37°C for 10 min and then centrifuged at room temperature at for 2 min. The

upper phase (aqueous) was transferred to a new tube. The lower phase (detergent) was

washed with lysis buffer lacking detergent, and separation was performed as described

above. Following the wash, the lower phase was diluted into buffer lacking detergent.

Protein concentrations were determined by the BCA method and Rab6

geranylgeranylation was assessed by gel electrophoresis and Western blot.

Mineralization assay. Primary calvarial osteoblasts were plated in 24-well plates.

Treatment began when cells reached confluency. After 14 days of culture, cells were

fixed for one hour in ice-cold 70% ethanol. Cells were then washed thoroughly with

deionized water; mineralization was detected with 40 mM Alizarin red, pH =4.2 for 15

min. Following staining, cells were washed thoroughly with deionized water to remove

non-specific Alizarin red. Images were captured using a Canon EOS Rebel XS.

Mineralization was quantified as described previously(70). Briefly, Alizarin Red was

eluted with 10% acetic acid per well. Plates were shaken on an orbital rotator for 15

minutes. The acetic acid and cells were transferred to a 1.5 mL tube and heated to 85oC

for 10 minutes. The samples were then cooled on ice and centrifuged. The supernatant

was transferred to a new tube and 10% ammonium hydroxide was added. Samples were

vortexed vigorously and aliquots were transferred to a 96 well plate. The absorbance was

read at 405 nm on a Thermomax Microplate Reader (Molecular Devices). Absorbance

was compared to a standard curve; values were expressed as total mol Alizarin Red per

well.

Real-time quantitative polymerase chain reaction (qPCR). Primary calvarial

osteoblasts or primary BMSCs were plated in 6-well plates. Alternatively, primary

65

BMSCs cultured in adipogenic medium were plated in 10 cm plates. Treatment was

applied when cells reached confluency. Five days (calvarial osteoblasts) or seven days

(BMSCs) following the onset of treatment with GGPP or GGPP vehicle, total RNA was

isolated from each well using Qiashredders and the RNeasy Mini Kit (Qiagen).

Alternatively, BMSCs cultured with digeranyl bisphosphonate (DGBP) were cultured for

two weeks prior to RNA isolation. During the isolation, a DNase step was performed

(Qiagen). One µg of RNA was reverse-transcribed into cDNA using the iScript cDNA

Synthesis Kit (Bio-Rad). Real-time PCR was performed with Sybr Green Master Mix

(Applied Biosystems) on ABI SDS 7900 HT (Applied Biosystems). The real-time

protocol consisted of 2 minutes at 50 oC, 10 minutes at 95 oC, followed by 40 cycles of

95 oC (15 seconds) and 60 oC (one minute). Dissociation curves were obtained following

real-time qPCR to ensure the proper amplification of target cDNAs. Primers were

obtained from Integrated DNA Technologies (Iowa City, IA, USA) and eluted in TE

Buffer (Ambion). The primers for qPCR are found in Table 1.

Oil Red O-Staining. Primary calvarial osteoblasts or BMSCs were plated in 12-

well plates. Cells were treated in specified conditions for 14 days, washed twice with

1xPBS, and fixed in 4% paraformaldehyde for 30 min. Cells were washed twice with

PBS, followed by a wash with 2-propanol:deionized water (6:4). A 0.5% Oil Red O stock

was prepared in 2-propanol. The stock was mixed with DI H2O at a ratio of 6:4 Oil Red

O stock to DI H2O. This mixture was allowed to sit ten minutes and then filtered.

Following the wash steps, cells were stained with filtered Oil Red O for one hour. Oil

Red O was removed and cells were washed with DI H2O. Stained cells were stored in 1x

66

PBS until quantification. Staining was visualized with light microscopy and Oil Red O

positive cells were quantified.

Glucose Transport. Primary calvarial osteoblasts and MC3T3-E1 pre-osteoblasts

were plated in 6-well plates. Upon confluency, cells were serum starved overnight. Cells

were treated with control, GGPP vehicle, or 20 M GGPP for 4.5 h. Following treatment,

cells were washed with ice cold glucose transport buffer (140 mM NaCl, 5 mM KCl, 1

mM MgCl2, 1.5 mM CaCl2, 15 mM N-2-hydroxyethylpiperazine-N-2-ethanesulfonic

acid). Uptake experiments were performed as described previously(90). Briefly, the cells

were incubated with glucose transport buffer containing 0.1 mM cold D-glucose and 2-

deoxy-D-2-[1-3H] glucose (1 Ci/mL, ARC) for ten min at room temperature. The buffer

was removed and the cells were washed with glucose transport buffer. The cells were

lysed in 1 mL of 2% SDS lysis buffer. The radioactivity in each sample was measured by

a liquid scintillation counter. Glucose uptake was normalized to protein content as

measured by the BCA assay.

Statistical Analysis. Data are expressed as means standard error of the mean

(SEM). All experiments were repeated at least twice with similar results. Differences

between two groups were compared using unpaired student’s t tests. Statistical

significance was defined by p values less than 0.05.

Results

GGPP inhibits osteoblast differentiation. To determine the effect of GGPP on

osteoblast differentiation, we treated primary calvarial osteoblasts with osteoblastic

differentiation medium in the presence or absence of GGPP. As shown in Figure 14A and

67

B, 10 and 20 M GGPP significantly inhibited osteoblast mineralization, as measured by

Alizarin red staining. Consistent with this, expression of the osteoblastic genes ALP,

BMP-2, and OCN were significantly decreased following five days of treatment with 10

or 20 M GGPP (Figure 14C). In contrast, there was no significant effect of GGPP on the

expression of Col1a1.

GGPP induces expression of PPARγ. Since it has been noted that GGPPS

expression increases during the adipogenic differentiation of 3T3-L1 cells(60), we

measured whether exogenous GGPP could stimulate the expression of PPARγ, a

dominant regulator of adipogenesis(88). Interestingly, 20 µM GGPP led to a significant

increase in total PPARγ expression in primary calvarial osteoblasts (Figure 15A).

Specific measurement of the PPAR2 splice variant, which is predominantly expressed

by adipocytes, showed that, similar to PPARγ expression, 20 M GGPP significantly

increased the expression of PPARγ2 as compared to control and GGPP vehicle-treated

calvarial osteoblasts (Figure 15A).

The effect of GGPP on total PPARγ and PPAR2 expression was also assessed in

BMSC cultures. As demonstrated in Figure 15B, treatment of BMSC cultures with 20

M GGPP for seven days led to a significant increase in PPARγ expression. This was

exhibited in adipogenic, control, and osteogenic culture conditions. Expression of PPAR

in GGPP-treated control or osteoblastic BMSC cultures was comparable to PPAR

expression levels in control adipogenic cultures. Similar to primary calvarial osteoblast

cultures, GGPP treatment increased the expression of the PPARγ2 splice variant in

adipogenic, control, and osteogenic conditions (Figure 15C).

68

Inhibition of PPARγ transcriptional activity does not prevent the effects of GGPP

on osteoblast differentiation. Agents that activate PPARγ, such as the thiazolidinediones,

are known to antagonize osteoblast differentiation (58). To test whether PPAR

activation was the mechanism by which GGPP inhibits osteoblast differentiation, we

employed the PPARγ antagonist GW9662. Primary calvarial osteoblasts were treated

with 20 µM GGPP or GGPP vehicle in the presence or absence of 0.1, 1.0, or 10 µM

GW9662. The highest concentration of GW9662 had a slight positive effect on osteoblast

matrix mineralization as measured by Alizarin red staining. However, GW9662 did not

prevent the effects of GGPP on the inhibition of osteoblast mineralization (Figure 16A

and B).

GGPP promotes adipogenesis. Because GGPP increased PPARγ expression, we

hypothesized that GGPP may stimulate adipogenesis. Primary calvarial osteoblasts were

cultured in osteoblastic differentiation medium for two weeks in the presence or absence

of 20 M GGPP. Cultures were stained with Oil Red O to detect lipid droplet containing

cells. Total Oil Red O-positive cells were quantified in each well using light microscopy.

As shown in Figure 17A, treatment with GGPP led to a significant increase in the number

of Oil Red O-positive cells per well. These experiments were also carried out in the

presence of 1.0 M rosiglitazone, a PPARγ agonist. At this concentration, rosiglitazone

had no significant effect on the number of Oil Red O-positive cells in control conditions.

The addition of rosiglitazone in the presence of GGPP led to a greater increase in the

number of Oil Red O-positive cells as compared to GGPP alone; this increase was not

significant. No Oil Red O-positive cells were detected in BMSC osteoblast cultures with

control, GGPP vehicle, or GGPP (data not shown). The addition of 1.0 µM rosiglitazone

69

led to the appearance of Oil Red O-positive cells in BMSC osteoblast cultures. 20 µM

GGPP significantly increased the formation of Oil Red O-positive cells in the presence of

rosiglitazone (Figure 17B).

The effect of GGPP on adipogenic gene expression was also assessed. GGPP

treatment of primary calvarial osteoblasts led to a small but significant increase in

adiponectin expression (Figure 17D). Treatment with rosiglitazone significantly

increased the expression of adiponectin as compared to control osteoblasts as well as

GGPP-treated osteoblasts. The combination of 1.0 µM rosiglitazone with GGPP led to a

significant increase in adiponectin expression as compared to rosiglitazone control or

GGPP alone. Also of note, GGPP treatment led to the increased expression of uncoupling

protein (UCP)-1 and to slightly decreased glucose transporter (Glut)-4 expression (Figure

17E and F). Co-treatment with rosiglitazone did not result in a significant increase in

PPARγ (Figure 17C) or UCP1 (Figure 17E). Similarly, co-treatment with rosiglitazone

did not potentiate the decrease in Glut4 expression (Fig. 17F).

Specific depletion of GGPP inhibits adipogenesis. Since GGPP increased the

expression of certain adipogenic markers and the number of Oil Red O-positive cells, we

tested whether specific depletion of GGPP would inhibit adipocyte differentiation. As

described in Chapter II, DGBP is a specific inhibitor of GGPPS (64, 65). Treatment of

BMSCs for 24 h with increasing concentrations of DGBP led to a significant decrease in

intracellular GGPP. Similar to the treatment of MC3T3-E1 pre-osteoblasts and primary

calvarial osteoblasts, DGBP treatment resulted in a significant increase in the GGPPS

substrate FPP (Figure 18A)(85).

70

BMSCs were treated with adipogenic medium in the presence or absence of

DGBP (0.1, 0.5, or 1.0M) for two weeks, followed by Oil Red O-staining. Alternatively,

BMSCs were treated with 1.0 M lovastatin. As expected, lovastatin reduced the

formation of Oil Red O positive-cells. Similarly, DGBP reduced formation of Oil Red O-

positive cells at concentrations of 0.5 and 1.0 M (Figure 18B and C). Adipogenic

differentiation was also assessed by expression of adipogenic markers. Similar to

treatment with lovastatin, treatment with 1.0 M DGBP inhibited the expression of the

adipogenic genes PPAR, adiponectin, and CCAAT/enhancer-binding protein (C/EBP)

(Figure 18D).

GGPP treatment does not increase protein geranylgeranylation. Previous reports

have suggested that exogenous GGPP causes cellular effects through an increase in

protein geranylgeranylation. To determine whether the addition of GGPP increases

geranylgeranylation in osteoblasts we analyzed the detergent/aqueous fractionation of the

geranylgeranylated protein Rab6. Rab6 is a GGTase II substrate. Prenylated Rab6

associates with the detergent fraction whereas, unprenylated Rab6 localizes to the

aqueous fraction. MC3T3-E1 pre-osteoblasts were treated for 24 h with 20 M lovastatin,

GGPP vehicle, 10 M GGPP, or 20 M GGPP. In control conditions Rab6 was detected

in the detergent phase and not the aqueous phase indicating that all of the Rab6 was

prenylated (Figure 19). As expected, treatment of the cells with lovastatin, which depletes

GGPP, decreased the levels of Rab6 in the detergent phase and caused the appearance of

unprenylated Rab6 in the aqueous phase. The addition of GGPP vehicle, 10 M GGPP or

20 M GGPP did not result in detection of aqueous Rab6. 20 M GGPP caused a slight

reduction in Rab6 protein levels as evidenced by decreased Rab6 in the detergent phase

71

with no appearance of Rab6 in the aqueous phase. These results suggest that exogenous

GGPP does not increase protein geranylgeranylation of Rab6.

GGPP decreases insulin receptor activation in osteoblasts. A report by Chen, et

al. demonstrated that GGPP could activate PTPase-1B activity in vitro, leading to

enhanced dephosphorylation of an insulin receptor (InsR) peptide substrate(91). Recent

studies have demonstrated that InsR signaling is essential for osteoblast differentiation

(92, 93). To test whether GGPP inhibits osteoblast differentiation through activation of

PTPase-1B and negative regulation of InsR signaling, the effect of GGPP on InsR

phosphorylation was tested. MC3T3-E1 pre-osteoblasts were pretreated with 20 M

GGPP or vehicle for 30 min, followed by stimulation with 100 nM insulin. GGPP pre-

treatment decreased the phosphorylation of the InsR at Tyr 1150/1151 in response to the

insulin treatment at each time point tested (Figure 20A). Since decreased InsR signaling

in InsR knockout osteoblasts is associated with decreased Runx2 activity and the

increased expression of the runx2 inhibitors, twist1 and twist2(93), we tested whether

similar effects occurred in osteoblasts treated with GGPP. GGPP treatment led to a

decrease in runx2 expression; however this was not significant (Figure 20B). While

twist1 expression was not significantly altered in primary calvarial osteoblasts cultured

with GGPP for 5 days, twist2 expression was increased approximately four-fold (Figure

20C).

GGPP increases insulin signaling. A recent study noted that GGPPS expression

leads to enhanced Erk1/2 activation and subsequent negative regulation of insulin

receptor substrate (IRS)-1(94). We therefore examined the effect of exogenous GGPP on

Erk1/2 phosphorylation. We found that Erk1/2 phosphorylation is increased with GGPP

72

treatment alone. This activation was increased further with insulin treatment to a greater

extent than cells treated with GGPP vehicle (Figure 21A). Phosphorylation of Akt was

not affected by GGPP. There was a slight increase in the phosphorylation of IRS-1 at

serine 612. In contrast to what would be expected by decreased InsR phosphorylation and

enhanced negative regulation of insulin signaling, glucose uptake by MC3T3-E1 pre-

osteoblasts and calvarial osteoblasts was enhanced by treatment with GGPP (Figure

21B).

Discussion

Herein, we demonstrate that GGPP inhibits osteoblast differentiation, and

enhances adipogenic differentiation of primary calvarial and BMSC osteoblast cultures.

Adipogenic differentiation of BMSCs could be inhibited through the specific inhibition

of GGPPS. GGPP pre-treatment decreased the phosphorylation of the InsR, enhanced

Erk1/2 phosphorylation, and increased glucose uptake, suggesting potential mechanisms

by which GGPP acts to inhibit osteoblast differentiation.

Our finding that GGPP negatively regulates osteoblast differentiation is consistent

with the conclusions of Yoshida, et al. that GGPP depletion is important for osteoblast

differentiation (55). However, we did not detect an effect for GGOH on the

differentiation of primary calvarial osteoblasts (data not shown). This may be due to our

use of primary osteoblasts instead of the MC3T3-E1 pre-osteoblast cell line utilized in

the study by Yoshida, et al. It is possible that the primary cells do not express the kinases

necessary to phosphorylate GGOH in the absence of GGPP-lowering agents. Due to the

lack of identification of these kinases, we were unable to test this hypothesis. Consistent

73

with a negative role for GGPP in osteoblast differentiation, we demonstrated in Chapters

II and III that GGPP levels decrease during osteoblast differentiation(84, 85).

Additionally, PTH treatment, which stimulates anabolic bone formation, has been shown

to reduce the expression of the upstream isoprenoid pathway enzyme mevalonate kinase.

The effect of PTH to stimulate BMP-2 expression was prevented by the addition of

exogenous GGPP, suggesting that PTH stimulates BMP-2 expression through the

negative regulation of mevalonate kinase expression and the resulting decrease in GGPP

(56).

GGPP treatment led to an induction of PPARγ expression. PPARγ exists as two

splice variants, PPARγ1 and PPARγ2. PPARγ1 is expressed in a wide range of tissues,

including the liver, adipose tissue, and bone, whereas PPARγ2 is expressed primarily in

adipogenic cells (87). GGPP treatment increased total PPAR as well as the PPAR2

splice variant expression. This presented the possibility that PPARγ activity was

responsible for the GGPP-mediated suppression of osteoblast differentiation, since agents

that activate PPAR have previously been noted for inhibition of osteoblast

differentiation (2, 35, 58). However, treatment with GW9662, an irreversible PPARγ

antagonist, did not prevent the effects of GGPP on osteoblast mineralization. GW9662

acts by covalently binding cysteine residue 285 on PPAR resulting in the loss of ligand

binding (95). While this experiment confirmed that the effect of GGPP to inhibit

osteoblast differentiation is not due to the induction of PPARγ transcriptional activity, it

remains possible that PPARγ may be inhibiting osteoblast differentiation through non-

transcriptional effects. This was not specifically tested in this study.

74

Age associated bone loss is accompanied by an increase in bone marrow adiposity

(2, 86). This has been described as a change in the intrinsic differentiation potential of

mesenchymal stem cells, increasing their tendency to differentiate towards the adipogenic

lineage (35). Interestingly, PPAR2 is upregulated in bone marrow from old animals, as

compared to that from adult animals (86). Due to the effect of GGPP to enhance PPARγ

expression and adipogenesis, it would be of great interest to determine whether GGPP

levels change during aging and potentially contribute to increased marrow adiposity.

As demonstrated in the recent publications by Ferron, et al. and Fulzele, et al.

osteoblasts express the InsR and InsR signaling in osteoblasts is necessary for osteoblast

differentiation and bone formation (92, 93). Fulzele, et al. found that InsR signaling

negatively regulates the expression of twist1 and 2, inhibitors of Runx2 transcriptional

activity. Therefore, InsR signaling led to enhanced Runx2 activity (93). The InsR is

negatively regulated by PTPase-1B in osteoblasts (92). A study by Chen, et al.

demonstrated that GGPP enhanced the activity of PTPase-1B in vitro(91). We

demonstrate that GGPP pre-treatment reduces the phosophorylation of the InsR.

Additionally, osteoblasts treated with GGPP display increased expression of the Runx2

inhibitor, twist2, consistent with data from Fulzele, et al. demonstrating negative

regulation of twist2 expression by InsR signaling. Together this suggests that increased

levels of GGPP may lead to decreased osteoblast differentiation through the inhibition of

InsR signaling (Figure 22A).

InsR signaling in osteoblasts has been demonstrated to be necessary for whole-

body glucose homeostasis due to its positive effect on OCN production by the osteoblast

(92, 93). Secreted OCN enhances -cell proliferation and insulin secretion, insulin

75

sensitivity, and energy expenditure (96). Aged mice deficient for InsR expression in

osteoblasts developed increased peripheral adiposity, hyperglycemia, and insulin

resistance (93). Increased levels of GGPP may additionally lead to reduced serum insulin,

hyperglycemia, and insulin resistance due to reduced expression of OCN by osteoblasts

through the negative regulation of insulin receptor signaling in osteoblasts.

In addition to the effects specifically on osteoblasts, PTPase-1B is associated with

the responsiveness of several other tissues to insulin. Because of this, PTPase1B

knockout mice display increased insulin sensitivity and are protected from diet-induced

obesity. In contrast, PTPase-1B overexpression results in insulin resistance due to

reduced insulin receptor signaling (97). The data presented here suggests the possibility

that, in addition to its effects on osteoblasts, GGPP levels may play a role in pathological

insulin resistance and diabetes.

Interestingly, a recent study demonstrates that GGPPS expression is associated

with increased Erk activation, leading to negative regulation of the InsR substrate, IRS-

1(94). We demonstrate here that GGPP treatment increased Erk1/2 phosphorylation. We

also noted a slight increase in the phosphorylation of IRS-1 at serine 612.

Phosphorylation at the residue negatively affects IRS-1 signaling. This suggests that in

addition to GGPP inhibiting InsR signaling through enhanced PTPase-1B activity, GGPP

may also inhibit InsR signaling through Erk activation and the subsequent inhibition of

the InsR substrate, IRS-1 (Figure 22A).

Contrary to what would be expected with decreased InsR phosphorylation and

increased Erk-mediated negative regulation of insulin signaling, glucose transport was

increased by GGPP treatment of MC3T3-E1 pre-osteoblasts and primary calvarial

76

osteoblasts. This evidence provides yet another possible mechanism by which GGPP

inhibits osteoblast differentiation and promotes adipogenesis. GGPP has been shown to

antagonize LXR(26). LXR is known to be involved in hepatic insulin signaling(98) and

LXR/b knockout mice exhibit enhanced glucose uptake in vivo(99). Inhibition of LXR

by GGPP may explain the increased glucose uptake in treated osteoblasts. Additionally,

certain oxysterols are agonists for LXR; oxysterols have been shown to stimulate

osteoblast differentiation and impair adipogenesis in vitro(100). This suggests that GGPP

may mediate its effects on osteoblasts through antagonism of LXR (Figure 22B).

Although endogenous GGPP levels have not been thoroughly studied in

conditions of aging or disease, one recent publication demonstrated a link between a

single nucleotide polymorphism (SNP) in GGPPS and BMD. In a population of Korean

women, patients with a homozygous deletion allele at -8188 of GGPPS (rs3840452)

displayed higher BMD at the femoral neck than women heterozygous and homozygous

for the insertion at this SNP. The response rate of women exhibiting the homozygous

deletion allele in GGPPS to BP therapy was lower; these patients displayed a seven-fold

higher risk of non-response to BP therapy (63). While the consequence this SNP has not

been studied, the data presented here leads to the speculation that the deletion allele may

lead to decreased GGPPS expression and that the decreased GGPP levels subsequently

lead to higher BMD. The population size in the evaluation of this SNP was very low, and

these results must be verified in a larger population. While the aim of their study was to

determine the effects of the GGPPS variant on osteoclasts, it cannot be ruled out that this

SNP may also influence osteoblast differentiation and bone formation. The study by

77

Choi, et al. provides preliminary evidence that in vivo GGPP levels have a role in

regulating the skeleton.

Altogether, our investigations demonstrate that GGPP negatively regulates

osteoblast differentiation resulting in increased PPARγ expression and enhanced

adipogenesis. Also, our results demonstrating decreased InsR phosphorylation, enhanced

Erk1/2 phosphorylation, and increased glucose transport in response to GGPP treatment

have intriguing implications for a role of GGPP in the modulation of energy metabolism

as well as skeletal homeostasis. The mechanism by which GGPP exerts these effects

should be further characterized in order to determine novel therapeutic targets for the

treatment of diseases such as osteoporosis and diabetes.

78

Figure 14. GGPP inhibits primary calvarial osteoblast differentiation. A and B, primary calvarial osteoblasts were treated with 5-20 µM GGPP for two weeks. A, Mineralization was assessed by Alizarin red staining followed by (B) elution and quantification of the dye at 405 nm. Data are expressed as mole Alizarin red per well, percent vehicle (mean ± SEM), *p<0.05. C, Primary calvarial osteoblasts were treated with 10 or 20 µM GGPP for five days. Osteoblastic gene expression was assessed by qPCR. Expression of osteoblastic genes was normalized to the expression of GAPDH, and expressed as percent vehicle (mean±SEM), *p<0.05, n=3.

79

Figure 15. GGPP increases the expression of PPARγ. A, primary calvarial osteoblasts were treated with GGPP for 5 days. Expression of total PPARγ and the PPARγ2 splice variant was assessed by qPCR. B and C, Primary BMSCs were treated in adipogenic, control, or osteogenic medium for seven days in the presence or absence of 20 µM GGPP. qPCR was used to assess the expression of (B) total PPARγ and (C) PPARγ2. Expression was normalized to GAPDH. Data are expressed as relative units (mean ± SEM), *p<0.05 as compared to GGPP vehicle in each condition, n=3.

80

Figure 16. Inhibition of PPARγ transcriptional activity does not prevent the effects of GGPP on osteoblast mineralization. Primary calvarial osteoblasts were treated with increasing concentrations of GW9662 in the presence or absence of 20 µM GGPP. Mineralization was assessed by (A) Alizarin red staining followed by (B) elution and quantification of the dye at 405 nm. Data are expressed as percent vehicle (mean ± SEM) n=3.

81

Figure 17. GGPP enhances adipogenesis. A, Primary calvarial osteoblasts were treated with 20 µM GGPP for two weeks in the presence or absence of 1 µM rosiglitazone. Total Oil Red O-positive cells per well were quantified (mean SEM). ap<0.05 as compared to vehicle-treated osteoblasts, bp<0.05 as compared to rosiglitazone treated osteoblasts, n=3. B, BMSC osteoblast cultures were treated with 1 M rosiglitzaone in the presence of control, GGPP vehicle, or 20 M GGPP for two weeks. The number of Oil Red O-positive cells were quantified in 5 visual fields per well (mean SEM), *p<0.05 as compared to GGPP vehicle, n=3. C-F, Primary calvarial osteoblasts were treated with control, GGPP vehicle, or 20 M GGPP for five days. qPCR was used to assess expression of (C) PPAR, (D) adiponectin, (E) UCP1, and (F) Glut4. Expression was normalized to GAPDH. Data are expressed as relative units (mean S EM). ap<0.05 as compared to vehicle-treated osteoblasts, bp<0.05 as compared to rosiglitazone-treated osteoblasts, cp<0.05 as compared to GGPP treated osteoblasts, n=3.

82

83

Figure 18. Specific inhibition of GGPPS decreases adipogenesis. A, BMSCs were treated with DGBP (0.1-100 µM) for 24 h. DGBP concentration-dependently reduced BMSC GGPP and increased the substrate FPP levels, n=2. B-D, BMSCs were treated in adipogenic medium with 1 µM Lov or 0.1-1.0 µM DGBP for two weeks. B, Cultures were stained with Oil Red O and (C) Oil Red O-positive cells were quantified in 5 visual fields per well and averaged. *p<0.05, n=3. D, mRNA was isolated and qPCR was used to analyze the expression of adipogenic genes. Expression was normalized to GAPDH and data are expressed as relative units (mean SEM). *p<0.05 as compared to control treatment, n=3.

84

Figure 19. GGPP treatment does not increase geranylgeranylation. MC3T3-E1 pre-osteoblasts were treated with indicated treatments for 24 h. Following lysis, the cell lysates were fractionated to aqueous and detergent phases. Western blot was used to assess the amounts of geranylgeranylated (detergent fraction) and non-geranylgeranylated (aqueous fraction) Rab6.

85

Figure 20. GGPP reduces InsR phosphorylation. A, Serum starved MC3T3-E1 pre-osteoblasts were pre-treated 30 min with GGPP or vehicle, followed by treatment with 100 nM insulin for 0-30 min. Western blots were used to detect phosphorylated InsR (Tyr1150/1151), total InsR, and α-tubulin (loading control). B and C, primary calvarial osteoblasts were treated with GGPP for 5 days. qPCR was used to assess expression of (B) Runx2 and (C) Twist1 and Twist2. mRNA was normalized to GAPDH mRNA (mean ± SEM). *p<0.05, n=3.

86

Figure 21. GGPP enhances insulin-induced Erk1/2 activation and glucose uptake. A, Serum starved primary calvarial osteoblasts (left) and MC3T3-E1 pre-osteoblasts (right) were pre-treated with GGPP for 30 min followed by treatment with 100 nM insulin for 0-30 min. Western blots were used to detect phosphorylated Erk1/2, total Erk1/2, phosphorylated Akt, total Akt, phosphorylated IRS-1 (serine 612), and total IRS-1. B. Serum starved primary calvarial osteoblasts (left) and MC3T3-E1 pre-osteoblasts (right) were treated with 20 µM GGPP for 4.5 h. Following the treatment, glucose uptake was measured. Data are expressed as disintegrations per minute (DPM) per g protein, mean ± SEM.

87

Figure 22. Proposed mechanisms for the inhibition of osteoblast differentiation by GGPP.

88

CHAPTER V: SUMMARY

The Isoprenoid Pathway and Osteoblast Differentiation

Based on published findings that statins stimulate osteoblast differentiation and

bone formation(36-40, 42, 67, 101, 102) and that this occurs through the depletion of

GGPP(41, 43), we hypothesized that specific depletion of GGPP with DGBP would

similarly lead to osteoblast differentiation. In contrast to our hypothesis, DGBP treatment

inhibited osteoblast differentiation as measured by osteoblast gene expression and matrix

mineralization. This led us to explore the effects of FPP accumulation, a secondary effect

of GGPPS inhibition, on osteoblast differentiation.

To determine whether FPP accumulation inhibits osteoblast differentiation, we

employed the SQS inhibitor, ZGA. Inhibition of SQS with ZGA led to a greater

accumulation of FPP than can be achieved by GGPPS inhibition. ZGA treatment also led

to increases in GGPP; however GGPP did not accumulate to the extent of FPP.

Consistent with our hypothesis, ZGA inhibited the differentiation of osteoblasts as

measured by gene expression, ALP activity, and matrix mineralization. ZGA also

inhibited osteoblast viability, suggesting that the accumulation of FPP prevents osteoblast

expansion. Co-treatment of calvarial cells with ZGA and lovastatin prevented the

accumulation of FPP as well as the inhibitory effects of ZGA on osteoblast viability,

differentiation, and matrix mineralization. These inhibitory effects could be restored by

the addition of mevalonate, the product downstream of the statin target, HMGCR.

Together this mechanistic data suggests that FPP accumulation negatively regulates

osteoblast differentiation.

89

As mentioned previously, FPP has been shown to bind and activate several

nuclear hormone receptors, including the estrogen, thyroid, and glucocorticoid receptors

(27). Interestingly, upon analysis of the glucocorticoid receptor, we demonstrated that

DGBP-induced FPP accumulation led to the activation of the glucocorticoid receptor as

measured by nuclear translocation and phosphorylation(85). A role for glucocorticoid

receptor activation would be consistent with data in Chapter III showing that the negative

effects of FPP accumulation were not prevented by inhibition of farnesylation(84).

Glucocorticoids have long been known to have negative effects on the skeleton, therefore

the potential role for FPP accumulation in glucocorticoid receptor activation and

regulation of skeletal homeostasis is very intriguing. Although not tested in this study, it

also remains possible that FPP activates other nuclear hormone receptors in osteoblasts to

influence proliferation, differentiation and matrix mineralization.

Consistent with previous findings that GGPPS expression decreases during

MC3T3-E1 pre-osteoblast differentiation, intracellular FPP and GGPP levels were found

to decrease during the differentiation of both MC3T3-E1 pre-osteoblasts and primary rat

calvarial osteoblasts. In primary calvarial cells this was associated with decreased

expression of HMGCR and FPPS. In support of decreased isoprenoid pathway activity

during osteoblast differentiation, a study by Takase, et al. reported that the anabolic agent

PTH negatively regulates the expression of mevalonate kinase (56). As noted previously,

our finding of increased GGPPS expression with differentiation is in contrast to those of

Yoshida, et al, who demonstrated that GGPPS decreased during MC3T3-E1 pre-

osteoblast differentiation (55). It is possible that this difference is due to the use of a cell

line in contrast to primary cells. It is important to note, however, that our study and that

90

by Yoshida, et al. are in agreement that GGPP levels decrease during osteoblast

differentiation.

We also demonstrate here the negative effects of exogenous GGPP on primary

calvarial osteoblast differentiation. As mentioned previously, Yoshida, et al. published

the negative effect of GGOH on MC3T3-E1 pre-osteoblast differentiation (55). In

contrast to their results, GGOH did not affect osteoblast differentiation in our study. This

may be due to fact that GGOH must be phosphorylated by intracellular kinases, whose

activity and expression we speculate may be regulated by the presence of GGPP.

Therefore, in the absence of GGPP-depleting agents, GGOH may not be converted to

GGPP in primary osteoblasts.

Because GGPPS expression has previously been shown to increase during

adipogenesis (60) and adipogenesis is impaired by GGPP depletion (42, 59), we tested

the effect of GGPP on the expression of PPARγ, a dominant regulator of adipogenic

differentiation. Interestingly, PPARγ expression was significantly increased by GGPP in

both primary calvarial osteoblast and BMSC cultures. We found that expression of

PPARγ2, the splice variant expressed predominantly be adipocytes, was also significantly

increased by GGPP. GGPP-stimulated PPARγ expression resulted in the increased

formation of lipid droplet containing cells, a classic marker of adipocyte formation, as

well as increased the expression of adiponectin and UCP1. UCP1 is a marker of brown

adipose tissue (BAT). In contrast, GGPP treatment led to a slight decrease in the

expression of Glut4, a marker of white adipose tissue (WAT). This suggests a potential

role for GGPP levels in the regulation of BAT formation. Consistent with a role for

GGPP to enhance PPARγ expression and adipocyte differentiation, specific depletion of

91

GGPP with DGBP inhibited the formation of lipid droplet-containing cells and

expression of the adipogenic genes PPARγ and adiponectin. In determining the

mechanism by which GGPP inhibits osteoblast differentiation, we first sought to

determine whether its effects were due to activation of PPARγ transcriptional activity,

which has been noted to negatively regulate osteoblast differentiation. Activators of

PPARγ are known inhibitors of osteoblast differentiation and reduce bone mass in in vivo

models(2, 35, 58). GW9662 covalently binds cysteine 285 of PPARγ resulting in loss of

its ligand binding and transcriptional activity(95). While GW9662 enhanced the matrix

mineralization of primary calvarial cells alone, co-treatment with GGPP did not prevent

the negative effects of GGPP on the inhibition of matrix mineralization. This suggests

that the effect of GGPP to inhibit osteoblast differentiation is independent of the effect of

GGPP on PPARγ activity. However, it cannot be ruled out that PPARγ may inhibit

osteoblast differentiation through a non-transcriptional mechanism.

There has been much interest recently in the role of InsR signaling in osteoblasts.

Ferron, et al. and Fulzele, et al. both demonstrated that osteoblasts deficient for the InsR

exhibit decreased differentiation and matrix mineralization(92, 93). In contrast, murine

osteoblasts deficient for osteotesticular (OST)-PTPase, the murine homolog of PTPase-

1B, exhibit enhanced osteoblast differentiation(96). Ferron, et al. show that the InsR is a

substrate for OST-PTP and PTPase-1B in osteoblasts, and OST-PTP knockout

osteoblasts exhibit higher levels of active, phosphorylated InsR (92). Interestingly,

PTPase-1B has been shown to be activated in vitro by GGPP(91), leading to enhanced

dephosphorylation of the InsR substrate. In our experiments, MC3T3-E1 pre-osteoblast

cultures pre-treated with GGPP displayed decreased phosphorylation of the InsR in

92

response to insulin treatment. Fulzele, et al. demonstrated that InsR signaling negatively

regulated the expression of twist1 and twist2, two negative regulators of Runx2 activity

(93). Consistent with a role for GGPP to negatively regulate the InsR in osteoblasts,

primary calvarial osteoblasts treated with GGPP exhibited increased expression of twist2

and decreased expression of Runx2. This suggests that GGPP may be inhibiting

osteoblast differentiation through the inhibition of InsR signaling.

A recent publication by Shen, et al. noted a correlation between GGPPS

expression, Erk activity, and insulin signaling. In their study, the authors demonstrated

that GGPPS expression led to increased Erk activation and subsequent inhibition of IRS-

1, a mediator of InsR signaling, suggesting a second mechanism by which GGPP

negatively regulates InsR signaling(94). Consistent with their work, we demonstrate

increased Erk1/2 activation in GGPP-treated osteoblasts. We also noted a slight increase

in phosphorylation of IRS-1 at serine 612, which negatively regulates IRS-1 signaling.

However in contrast to what would be expected, glucose uptake was increased by GGPP

treatment. One potential explanation for this effect may be antagonism of LXR by GGPP.

LXR has been reported to be involved in hepatic insulin signaling(98, 99), and LXR

knockout mice exhibit enhanced glucose uptake in response to insulin treatment(99).

Consistent with this as a potential mechanism explaining our results in osteoblasts, LXR

agonists stimulate osteoblast differentiation and inhibit adipogenesis in vitro. Altogether,

our results demonstrating decreased InsR phosphorylation, increased Erk activation, and

increased glucose uptake have important implications for skeletal bone formation as well

as energy homeostasis.

93

Future Studies

The work presented in this thesis presents many questions and potential future

directions. One question that is especially interesting is the potential role for GGPP to

enhancing the activity of PTPase-1B. First and foremost, it needs to be confirmed that the

effect of GGPP in decreasing InsR phosphorylation is indeed due to increased PTPase-1B

activity. This may be done by knocking down PTPase-1B in osteoblasts prior to the

GGPP and insulin treatments or a co-treatment with a PTPase-1B inhibitor. PTPase-1B

inhibitors are being developed for the treatment of diabetes, but are not currently

commercially available(103). If these studies confirm a role for GGPP in preventing InsR

phosphorylation through enhanced PTPase1B activity, experiments should be done to

confirm whether the mechanism of inhibiting osteoblast differentiation is through

negative regulation of InsR signaling. PTPase-1B regulates the activity of the InsR in

many tissues. It has been shown that mice deficient for PTPase1B exhibit enhanced

insulin sensitivity, whereas mice overexpressing PTPase1B display decreased insulin

sensitivity(97). Additionally, it should be determined whether activation of Erk by

exogenous GGPP leads to inhibition of IRS-1, which also may contribute to reduced InsR

signaling in osteoblasts. The role of GGPP to activate PTPase1B and Erk activity,

resulting in altered insulin sensitivity may potentially contribute to the pathogenesis of

diabetes and other metabolic disorders. It is also very important to further investigate the

role of LXR in the effects of GGPP on the osteoblast and adipocyte fate decision.

Potential experiments to confirm LXR inhibition as a mechanism for the negative effects

of GGPP on osteoblast differentiation and the enhanced glucose uptake include

overexpression of LXR.

94

While this study has shown very interesting roles for the isoprenoids FPP and

GGPP in regulating osteoblast differentiation, one is left wondering whether either of

these isoprenoids exists at concentrations necessary to exert the effects noted herein in

either physiological or pathological conditions. Several recent studies have reported

SNPs that affect the skeleton. Levy, et al. demonstrated that individuals with an AA

genotype at the FPPS SNP (rs2297480) displayed a significant increase in BMD at

several sites(61). These patients also exhibited an enhanced response to BP therapy(62).

A more recent study by Choi, et al. identified a deletion SNP in GGPPS (rs3840452)

which correlated with increased BMD at the femoral head. Patients with this genotype

had a greater chance of non-response to BP therapy(63). It is possible that these

polymorphisms may lead to higher or lower endogenous levels of FPP or GGPP.

Determining the functionality of these SNPs would be of great interest in determining

whether altered levels of FPP or GGPP contribute to the skeletal phenotypes observed in

these individuals.

Another interesting question lies in the field of aging. Age related bone loss is

associated with an increase in marrow adiposity(2, 6, 58). As mentioned previously,

osteoblasts and adipocytes are derived from mesenchymal stem cells(2, 6, 57).

Differentiation into either cell type is controlled by lineage specific transcription factor

expression(57). Interestingly, signals that promote osteoblastogenesis inhibit

adipogenesis, and vice versa. For example, agonists of PPARγ, such as the

thiazolidinediones, have been shown to inhibit osteoblast differentiation and bone

formation(6, 57, 58). Interestingly, the age-associated increase in marrow adiposity

correlates with increased PPARγ expression in the bone marrow(86), suggesting a shift in

95

differentiation potential with age (35). Together with data presented in this thesis, it

would be very interesting to investigate whether GGPP, which we have shown herein

increases PPARγ expression, may also increase during aging. This may also be

interesting in investigating other aging-related diseases, such as diabetes, in which cells

display an aging phenotype.

Conclusion

Altogether, the data presented in this thesis demonstrate roles for the isoprenoids

FPP and GGPP to negatively regulate osteoblast differentiation. This is consistent with

data from two cell models that show that these isoprenoids decrease during osteoblast

differentiation. The ability of FPP to activate nuclear hormone receptors in physiological

and pathological skeletal conditions should be further investigated. Additionally, in the

case of GGPP, we demonstrate that this isoprenoid plays a role in the osteoblast and

adipocyte fate decision through inhibition of osteoblast differentiation and the resulting

upregulation of the adipocytic transcription factor PPARγ. Therefore the potential role for

GGPP in metabolic physiology may be of great importance.

96

REFERENCES

1. Chau, J. F., Leong, W.F., Li, B., Signaling pathways governing osteoblast proliferation, differentiation and function. Histol Histopathol 2009, 24 (12), 1593-1606.

2. Muruganandan, S., Roman, A.A., Sinal, C.J, Adipocyte differentiation of bone marrow-derived mesenchymal stem cells: Cross talk with the osteoblastic program. Cell Mol Life Sci 2009, 66 236-253.

3. Canalis, E., Update in new anabolic therapies for osteoporosis. J Clin Endo Metab 2010, 95 (4), 1496-1504.

4. Gallagher, J. C., Sai, A.J., Molecular biology of bone remodeling: implications for new therapeutic targets for osteoporosis. Maturitas 2010, 65 301-307.

5. Baron, R., Ferrari, S., Russell, R.G.G., Denosumab and bisphosphonates: Different mechanisms of action and effects. Bone 2010,

6. de Paula, F., Horowitz, MC, Rosen, CJ, Novel insights into the relationship between diabetes and osteoporosis. Diabetes Metab Res Rev 2010, 26 622-630.

7. Kanis, J., WHO Study Group, Assessment of fracture risk and its application to screening for postmenopausal osteoporosis: synopsis of a WHO report. WHO Study Group. Osteoporosis Int 1994, 4 (6), 368-381.

8. Compston, J., Osteoporosis: social and economic impact. Radiol Clin N Am 2010, 48 (3), 477-482.

9. Clarke, B., Khosla, S, Physiology of bone loss. Radiol Clin N Am 2010, 48 (3), 483-495.

10. Burge, R., Dawson-Hughes, B, Solomon, DH, Wong, JB, King, A, Tosteson, A,

Incidence and economic burden of osteoporosis-related fractures in the United States, 2005-2025. J Bone Miner Res 2007, 22 (3), 465-75.

11. Close, P., Neuprez, A., Reginster, J.Y., Developments in the pharmacotherapeutic management of osteoporosis. Expert Opin Pharmacother 2006, 7 (12), 1603-1615.

12. Gong, L. A., R.B., Klein, T.E, Bisphosphonates pathway. Pharmacogenet Genomics 2010, 21 (1), 50-53.

13. Bergstrom, J., Bostedor, RG, Masarachia, PJ, Reszka, AA, Rodan, G, Alendronate is a specific, nanomolar inhibitor of farnesyl diphosphate synthase. Arch Biochem Biophys 2000, 373 (1), 231-41.

97

14. van Beek, E., Pieterman, E, Cohen, L, Lowik, C, Papapoulos, S, Farnesyl Pyrophosphate Synthase Is the Molecular Target of Nitrogen-Containing Bisphosphonates. Biochem Biophys Res Comm 1999, 264 (1), 108-111.

15. Luckman, S., Hughes, DE, Coxon, FP, Graham, R, Russell, G, Rogers, MJ, Nitrogen-containing bisphosphonates inhibit the mevalonate pathway and prevent post-translational prenylation of GTP-binding proteins, including Ras. J Bone Miner Res 1998, 13 (4), 581-589.

16. Monkkonen, H., Auriola, S, Lehenkari, P, Kellinsalmi, M, Hassinen, IE, Vepsalainen, J, Monkkonen, J , A new endogenous ATP analog (ApppI) inhibits the mitochondrial adenine nucleotide translocase (ANT) and is responsible for the apoptosis induced by nitrogen-containing bisphosphonates. Brit J Pharm 2006, 147 437-445.

17. Matsumoto, T., Nagase, Y, Yasui, T, Masuda, H, Nakamura, K, Tanaka, S, Segregation of Pro-apoptotic and Anti-resorptive Functions of Risedronate In Vivo. In J Bone Miner Res, Toronto, CA, 2010; Vol. 25.

18. Coleman, R., McCloskey, EV, Bisphosphonates in oncology. Bone 2011,

19. Girotra, M., Rubin, MR, Bilezikian, JP, Anabolic skeletal therapy for osteoporosis. Arq Bras Endocrinol Metab 2006, 50 (4), 745-754.

20. Goldstein, J., Brown, MS, Regulation of the mevalonate pathway. Nature 1990, 343 (6257), 425-430.

21. Holstein, S., Hohl, RJ, Isoprenoids: remarkable diversity of form and function. Lipids 2004, 39 (4), 293-309.

22. Perez-Sala, D., Protein isoprenylation in biology and disease: general overview and perspectives from studies with genetically engineered animals . Frontiers Biosci 2007 , 12 4456-4472.

23. Zhang, F. L., Casey, P.J., Protein Prenylation: Molecular Mechanisms and Functional Consequences. Annual Reviews Biochem 1996, 65 241-69.

24. Sebti, S., Protein farnesylation: Implications for normal physiology, malignant transformation, and cancer therapy. Cancer Cell 2005, 7 297-300.

25. Forman, B., Goode, E, Chen, J, Oro, AE, Bradley, DJ, Perlmann, T, Noonan, DJ, Burka, LT, McMorris, T, Lamph, WW, Evans, RM, Weinberger, C, Identification of a nuclear receptor that is activated by farnesol metabolites. Cell 1995, 81 (5), 687-693.

26. Forman, B., Ruan, B, Chen, J, Schroepfer, GJ Jr, Evans, RM, The orphan nuclear

receptor LXRalpha is positively and negatively regulated by distinct products of mevalonate metabolism. PNAS 1997, 94 (20), 10588-10593.

98

27. Das, S., Schapira, M, Tomic-Canic, M, Goyanka, R, Cardoza, T, Samuels, HH, Farnesyl pyrophosphate is a novel transcriptional activator for a subset of nuclear hormone receptors. Mol Endocrinology 2007, 21 (11), 2672-2686.

28. Vukelic, S., Stojadinovic, O, Pastar, I, Vouthounis, C, Krzyzanowska, A, Das, S,

Samuels, HH, Tomic-Canic, M, Farnesyl pyrophosphate inhibits epithelialization and wound healing through the glucocorticoid receptor. J Biol Chem 2010, 285 (3), 1980-1988.

29. Oh, D., Yoon, JM, Moon, MJ, Hwang, JI, Choe, H, Lee, JY, Kim, JI, Kim, S, Rhim, H, O'Dell, DK, Walker, JM, Na, HS, Lee, MG, Kwon, HB, Kim, K, Seong, JY, Identification of farnesyl pyrophosphate and N-arachidonylglycine as endogenous ligands for GPR92. J Biol Chem 2008, 283 (30), 21054-21064.

30. Liliom, K., Tsukahara, T, Tsukahara, R, Zelman-Femiak, M, Swiezewska, E, Tigyi, G, Farnesyl phosphates are endogenous ligands of lysophosphatidic acid receptors: inhibition of LPA GPCR and activation of PPARs. Biochimica et Biophysica Acta 2006, 1761 (12), 1506-1514.

31. Bang, S., Yoo, S, Yang, TJ, Cho, H, Hwang, SW, Farnesyl pyrophosphate is a novel pain-producing molecule via specific activation of TRPV3. J Biol Chem 2010, 285 (25), 19352-19371.

32. Zhou, Q., Liao, JK, Pleiotropic effects of statins. - Basic research and clinical perspectives-. Circ J 2010, 74 (5), 818-826.

33. Endo, A., Tsujita, Y, Kuroda, M, Tanzawa, K, Inhibition of cholesterol synthesis in vitro and in vivo by ML-236A and ML-236B, competitive inhibitors of 3-hydroxy-3-methylglutaryl-coenzyme A reductase. Eur J Biochem 1977, 77 (1), 31-6.

34. Toutouzas, K., Drakopoulou, M, Skoumas, I, Stefanadis, C, Advancing therapy for hypercholesterolemia. Expert Opin Pharmacother 2010, 11 (10), 1659-1672.

35. Moerman, E., Teng, K, Lipschitz, DA, Lecka-Czernik, B , Aging activates adipogenic and suppresses osteogenic programs in mesenchymal marrow stroma/stem cells: the role of PPAR gamma2 transcription factor and TGF-beta/BMP signaling pathways. Aging Cell 2004, 3 379-389.

36. Mundy, G., Garrett R, Harris, S, Chan, J, Chen, D, Rossini, G, Boyce, B, Zhao, M, Gutierrez, G, Stimulation of Bone Formation in Vitro and in Rodents by Statins. Science 1999, 286 (5446), 1946-1949.

37. Ruiz-Gaspa, S., Nogues, X, Enjuanes, A, Monllau, JC, Blanch, J, Carreras, R, Mellibovsky, L, Grinberg, D, Balcells, S, Diez-Perez, A, Pedro-Botet, J, Simvastatin and atorvastatin enhance gene expression of collagen type 1 and osteocalcin in primary human osteoblasts and MG-63 cultures. J Cell Biochem 2007, 101 (6), 1430-1438.

99

38. Chen, P., Sun, JS, Tsuang, YH, Chen, MH, Weng, PW, Lin, FH, Simvastatin promotes osteoblast viability and differentiation via Ras/Smad/Erk/BMP-2 signaling pathway. Nutrition Research 2010, 30 (3), 191-199.

39. Maeda, T., Matsunuma, A, Kawane, T, Horiuchi, N, Simvastatin promotes

osteoblast differentiation and mineralization in MC3T3-E1 cells. Biochem Biophys Research Comm 2001, 280 (3), 874-877.

40. Maeda, T., Kawane, T, Horiuchi, N, Statins augment vascular endothelial growth factor expression in osteoblastic cells via inhibition of protein prenylation. Endocrinology 2003, 144 (2), 681-692.

41. Maeda, T., Matsunuma, A, Kurahashi, I, Yanagawa, T, Yoshida, H, Horiuchi, N, Induction of osteoblast differentiation indices by statins in MC3T3-E1 cells. J Cell Biochem 2004, 92 (3), 458-471.

42. Li, X., Cui, Q, Kao, C, Wang, GJ, Balian, G, Lovastatin inhibits adipogenic and

stimulates osteogenic differentiation by suppressing PPARgamma2 and increasing Cbfa1/Runx2 expression in bone marrow mesenchymal cell cultures. Bone 2003, 33 (4), 652-659.

43. Ohnaka, K., Shimoda, S, Nawata, H, Shimokawa, H, Kaibuchi, K, Iwamoto, Y, Takayanagi, R, Pitavastatin enhanced BMP-2 and osteocalcin expression by inhibition of Rho-associated kinase in human osteoblasts. Biochem Biophys Res Comm 2001, 287 (2), 337-342.

44. Pagkalos J, C. J., Kang Y, Heliotis M, Tsiridis E, Mantalaris A, Simvastatin induces osteogenic differentiation of murine embryonic stem cells. J Bone Miner Res 2010,

45. Gutierrez GE, E. J., Garrett IR, Nyman JS, McCluskey B, Rossini G, Flores A, Neidre DB, Mundy GR, Transdermal lovastatin enhances fracture repair in rats. J Bone Miner Res 2008, 23 (11), 1722-30.

46. Skoglund B, F. C., Aspenberg P, Simvastatin improves fracture healing in mice. J Bone Miner Res 2002, 17 (11), 2004-8.

47. Chuengsamarn, S., Rattanamongkoulgul, S, Suwanwalaikorn, S, Wattanasirichaigoon, S, Kaufman, L, Effects of statins vs. non-statin lipid-lowering therapy on bone formation and bone mineral density biomarkers in patients with hyperlipidemia. Bone 2010, 46 (4), 1011-1015.

48. Tang, Q., Tran, GT, Gamie, Z, Graham, S, Tsialogiannis, E, Tsiridis, E, Linder, T, Tsiridis, E, Statins: under investigation for increasing bone mineral density and augmenting fracture healing. Expert Opin Investig Drugs 2008, 17 (10), 1435-1463.

100

49. Wang, P. S., Solomon, D.H., Mogun, H., Avorn, J., HMG-CoA reductase inhibitors and the risk of hip fractures in elderly patients. JAMA 2000, 283 (24), 3211-3216.

50. Uzzan, B., Cohen, R, Nicolas, P, Cucherat, M, Perret, GY, Effects of statins on

bone mineral density: a meta-analysis of clinical studies. Bone 2007, 40 (6), 1581-1587.

51. Duque, G., Rivas, D, Alendronate has an anabolic effect on bone through the

differentiation of mesenchymal stem cells. J Bone Miner Res 2007, 22 (10), 1603-1611.

52. Ebert, R., Zeck, S, Krug, R, Meissner-Weigl, J, Schneider, D, Seefried, L, Eulert, J, Jakob, F, Pulse treatment with zoledronic acid causes sustained commitment of bone marrow derived mesenchymal stem cells for osteogenic differentiation. Bone 2009, 44 (5), 858-864.

53. Orriss, I., Key, ML, Colston, KW, Arnett, TR, Inhibition of osteoblast function in vitro by aminobisphosphonates. J Cell Biochem 2009, 106 (1), 109-118.

54. Idris, A., Rojas, J, Greig, IR, Van't Hof, RJ, Ralston, SH, Aminobisphosphonates cause osteoblast apoptosis and inhibit bone nodule formation in vitro. Calcif Tissue Int 2008, 82 (3), 191-201.

55. Yoshida, T., Asanuma, M, Grossmann, L, Fuse, M, Shibata, T, Yonekawa, T, Tanaka, T, Ueno, K, Yasuda, T, Saito, Y, Tatsuno, I, Geranylgeranyl-pyrophosphate (GGPP) synthase is down-regulated during differentiation of osteoblastic cell line MC3T3-E1. FEBS Letters 2006, 580 (22), 5203-5207.

56. Takase, H., Yano, S, Yamaguchi, T, Kanazawa, I, Hayashi, K, Yamamoto, M, Yamauchi, M, Sugimoto, T, Parathyroid hormone upregulates BMP-2 mRNA expression through mevalonate kinase and Rho kinase inhibition in osteoblastic MC3T3-E1 cells. Horm Metab Res 2009, 41 (12), 861-865.

57. Tontonoz, P., Spiegelman, B.M, Fat and Beyond: The Diverse Biology of PPARgamma. Annu Rev Biochem 2008, 77 289-312.

58. Nuttall, M. E., Gimble, J.M., Controlling the balance between osteoblastogenesis and adipogenesis and the consequent therapeutic implications. Curr Op Pharm 2004, 4 290-294.

59. Song, C., Guo, Z, Ma, Q, Chen, Z, Liu, Z, Jia, H, Dang, G, Simvastatin induces osteoblastic differentiation and inhibits adipocytic differentiation in mouse bone marrow stromal cells. Biochem Biophys Res Comm 2003, 308 458-62.

101

60. Vicent, D., Maratos-Flier, E, Kahn, CR, The branch point enzyme of the mevalonate pathway for protein prenylation is overexpressed in the ob/ob mouse and induced by adipogenesis. Molecular and Cellular Biology 2000, 20 (6), 2158-2166.

61. Levy, M., Parker, RA, Ferrell, RE, Zmuda, JM, Greenspan, SL, Farnesyl diphosphate synthase: a novel genotype association with bone mineral density in elderly women. Maturitas 2007, 57 (3), 247-52.

62. Marini, F., Falchetti, A, Silvestri, S, Bagger, Y, Luzi, E, Tanini, A, Christiansen,

C, Brandi, ML, Modulatory effect of farnesyl pyrophosphate synthase (FDPS) rs2297480 polymorphism on the response to long-term amino-bisphosphonate treatment in postmenopausal osteoporosis. Current Medical Research and Opinion 2008, 24 (9), 2609-2615.

63. Choi, H., Choi, JY, Cho, SW, Kang, D, Han, KO, Kim, SW, Kim, WY, Chung, YS, Shin, CS , Genetic polymorphism of geranylgeranyl diphosphate synthase (GGSP1) predicts bone density response to bisphosphonate therapy in Korean women. Yonsei Med J 2010, 51 (2), 231-238.

64. Wiemer, A. J., Tong, H, Swanson, K.M., Hohl, R.J., Digeranyl bisphosphonate inhibits geranylgeranyl pyrophosphate synthase. Biochem Biophys Res Comm 2007, 353 (4), 921-925.

65. Shull, L., Wiemer, AJ, Hohl, RJ, Wiemer, DF, Synthesis and biological activity of isoprenoid bisphosphonates. Bioorg Med Chem 2006, 14 (12), 4130-4136.

66. Gutierrez, G., Lalka, D, Garrett, IR, Rossini, G, Mundy, GR, Transdermal

application of lovastatin to rats causes profound increases in bone formation and plasma concentrations. Osteoporosis Int 2006, 17 1033-42.

67. Skoglund, B., Forslund, C, Aspenberg, P, Simvastatin improves fracture healing in mice. J Bone Miner Res 2002, 17 (11), 2004-8.

68. Hayashi, K., Yamaguchi, T, Yano, S, Kanazawa, I, Yamauchi, M, Yamamoto, M, Sugimoto, T, BMP/Wnt antagonists are upregulated by dexamethasone in osteoblasts and reversed by alendronate and PTH: Potential therapeutic targets for glucocorticoid-induced osteoporosis. Biochem Biophys Res Comm 2009, 379 261-66.

69. Tong, H., Holstein, SA, Hohl, RJ, Simultaneous determination of farnesyl and geranylgeranyl pyrophosphate levels in cultured cells. Anal Biochem 2005, 336 (1), 51-59.

70. Gregory, C., Gunn, WG, Peister, A, Prockop, DJ, An Alizarin red-based assay of mineralization by adherent cells in culture: comparison with cetylpyridinium chloride extraction. Anal Biochem 2004, 329 (1), 77-84.

102

71. Nancollas, G., Tang, R, Phipps, RJ, Henneman, Z, Gulde, S, Wu, W, Mangood, A, Russell, RGG, Ebetino, FH, Novel insights into actions of bisphosphonates on bone: Differences in interactions with hydroxyapatite. Bone 2006, 38 617-27.

72. Canalis, E., Mazziotti, G, Giustina, A, Bilezikian, JP, Glucocorticoid-induced osteoporosis: pathophysiology and therapy. Osteoporos Int 2007, 18 1319-1328.

73. Sher, L., Harrison, JR, Adams, DJ, Kream, BE, Impaired cortical bone acquisition and osteoblast differentiation in mice with osteoblast-targeted disruption of glucocorticoid signaling. Calcif Tissue Int 2006, 79 (2), 118-125.

74. Endo, A., The discovery and development of HMG-CoA reductase inhibitors. J Lipid Res 1992, 33 (11), 1569-82.

75. vanBeek, Farnesyl Pyrophosphate Synthase Is the Molecular Target of Nitrogen-Containing Bisphosphonates. Biochem and Biophys Res Comm 1999, 264 (1), 108-111.

76. Reinholz, G., Getz, B, Pederson, L, Sanders, ES, Subramaniam, M, Ingle, JN,

Spelsberg, TC, Bisphosphonates Directly Regulate Cell Proliferation, Differentiation, and Gene Expression in Human Osteoblasts. Cancer Research 2000, 60 (21), 6001-6007.

77. Im, G., Qureshi, SA, Kenney, J, Rubash, HE, Shanbhag, AS, Osteoblast proliferation and maturation by bisphosphonates. Biomaterials 2004, 25 (18), 4105-15.

78. Kim, H., Kim, JH, Abbas, AA, Yoon, TR, Alendronate enhances osteogenic differentiation of bone marrow stromal cells: a preliminary study. Clin Orthop Relat Res 2009, 467 (12), 3121-3128.

79. Panzavolta, S., Torricelli, P, Bracci, B, Fini, M, Bigi, A, Alendronate and Pamidronate calcium phosphate bone cements: setting properties and in vitro response of osteoblast and osteoclast cells. Journal of Inorganic Biochem 2009, 103 (1), 101-106.

80. Fu, L., Tang, T, Miao, Y, Zhang, S, Qu, Z, Dai, K, Stimulation of osteogenic differentiation and inhibition of adipogenic differentiation in bone marrow stromal cells by alendronate via ERK and JNK activation. Bone 2008, 43 (1), 40-47.

81. Pozzi, S., Vallet, S, Mukherjee, S, Cirstea, D, Vaghela, N, Santo, L, Rosen, E, et al, High-dose zoledronic acid impacts bone remodeling with effects on osteoblastic lineage and bone mechanical properties. Clin Cancer Res 2009, 15 (18), 5829-39.

103

82. Kellinsalmi, M., Monkkonen, H, Monkkonen, J, Leskela, HV, Parikka, V, Hamalainen, M, Lehenkari, P, In vitro comparison of clodronate, pamidronate and zoledronic acid effects on rat osteoclasts and human stem cell-derived osteoblasts. Basic and Clinical Pharmacology and Toxicology 2005, 97 (6), 382-391.

83. Yoshida T, A. M., Grossmann L, Fuse M, Shibata T, Yonekawa T, Tanaka T,

Ueno K, Yasuda T, Saito Y, Tatsuno I, Geranylgeranyl-pyrophosphate (GGPP) synthase is down-regulated during differentiation of osteoblastic cell line MC3T3-E1. FEBS Letters 2006, 580 (22), 5203-5207.

84. Weivoda, M., Hohl, RJ, The Effects of Farnesyl Pyrophosphate Accumulation on

Calvarial Osteoblast Differentiation. Endocrinology 2011, In press

85. Weivoda, M., Hohl, RJ, The effects of direct inhibition of geranylgeranyl pyrophosphate synthase on osteoblast differentiation. J Cell Biochem 2011, 1506-1513.

86. Lecka-Czernik, B., Rosen, CJ, Kawai, M, Skeletal aging and the adipocyte program: New insights from an "old" molecule. Cell Cycle 2010, 9 (18), 3648-3654.

87. Kawai, M., Rosen, CJ, PPARgamma: a circadian transcription factor in adipogenesis and osteogenesis. Nat Rev Endocrinol 2010, 6 629-636.

88. Lecka-Czernik, B., Gubrij, I, Moerman, EJ, Kajkenova, O, Lipschitz, DA, Manolagas, SC, Jilka, RL, Inhibition of Osf2/Cbfa1 Expression and Terminal Osteoblast Differentiation by PPARgamma2. J Cell Biochem 1999, 74 357-371.

89. Bordier, C., Phase separation of integral membrane proteins in triton X-114 solution. J Biol Chem 1981, 256 1604-1607.

90. Zoidis, E., Ghirlanda-Keller, C, Schmid, C, Stimulation of glucose transport in osteoblastic cells by parathyroid hormone and insulin-like growth factor I. Mol Cell Biochem 2011, 348 33-42.

91. Chen, H., Chen, CH, Chuang, NN, Differential effects of prenyl pyrophosphates on the phosphatase activity of phosphotyrosyl protein phosphatase. J Exp Zool A Comp Exp Biol 2004, 301 (4), 307-316.

92. Ferron, M., Wei, J, Yoshizawa, T, Del Fattore, A, DePinho, RA, Teti, A, Ducy, P, Karsenty, G, Insulin signaling in osteoblasts integrates bone remodeling and energy metabolism. Cell 2010, 142 296-308.

93. Fulzele, K., Riddle, RC, DiGirolamo, DJ, Cao, X, Wan, C, Chen, D, Faugere, M, Aja, S, Hussain, MA, Bruning, JC, CLemens, TL, Insulin receptor signaling in osteoblasts regulates postnatal bone acquisition and body composition. Cell 2010, 142 309-319.

104

94. Shen, N., Yu, X, Pan, FY, Gao, X, Xue, B, Li, CJ, An early response transcription factor, Egr-1, enhances insulin resistance in type 2 diabetes with chronic hyperinsulinism. J Biol Chem 2011,

95. Leesnitzer, L., Parks, DJ, Bledsoe, RK, Cobb, JE, Collins, JL, Consler, TG, Davis, RG, Hull-Ryde, EA, Lenhard, JM, Patel, L, :lunket, KD, Shenk, JL, Stimmel, JB, Therapontos, C, Willson, TM, Blanchard, SG, Functional consequences of cysteine modification in the ligand binding sites of perxoisome proliferator activated receptors by GW9662. Biochem 2002, 41 6640-6650.

96. Lee, N., Sowa, H, Hinoi, E, Ferron, M, Ahn, JD, Confavreux, C, Dacquin, R, Mee, PJ, McKee, MD, Jung, DY, Zhang, Z, Kim, JK, Mauvais-Jarvis, F, Ducy, P, Karsenty, G, Endocrine Regulation of Energy Metabolism by the Skeleton. Cell 2008, 130 456-469.

97. Julien, S., Tremblay, ML, Insulin Receptor PTP:PTP1B. In Handbook of Cell Signaling, Elsevier: 2010; Vol. II, pp 811-815.

98. Basciano, H., Miller, A, Baker, C, Naples, M, Adeli, K, LXRalpha activation perturbs hepatic insulin signaling and stimulates production of apolipoprotein B-containing lipoproteins. Am J Physiol Gastrointest Liver Physiol 2009, 297 G323-G332.

99. Risan Tobin, K., Ulven, SM, Schuster, GU, Hermansen Steineger, H, Maehle Andresen, S, Gustafsson, J, Nebb, HI, Liver X Receptors as Insulin-mediating Factors in Fatty Acid and Cholesterol Biosynthesis. J Biol Chem 2002, 277 (12), 10691-10697.

100. Kha, H., Basseri, B, Shouhed, D, Richardson, J, Tetradis, S, Hahn, TJ, Parhami, F, Oxysterols Regulate Differentiation of Mesenchymal Stem Cells: Pro-Bone and Anti-Fat. J Bone Miner Res 2004, 19 (5), 830-840.

101. Pagkalos, J., Cha, JM, Kang, Y, Heliotis, M, Tsiridis, E, Mantalaris, A, Simvastatin induces osteogenic differentiation of murine embryonic stem cells. J Bone Miner Res 2010,

102. Gutierrez, G., Edwards, JR, Garrett, IR, Nyman, JS, McCluskey, B, Rossini, G, Flores, A, Neidre, DB, Mundy, GR, Transdermal lovastatin enhances fracture repair in rats. J Bone Miner Res 2008, 23 (11), 1722-30.

103. Mattila, E., Ivaska, J, High-Througput methods in identification of protein tyrosine phosphatase inhibitors and activators. Anticancer Agents Med Chem 2011.