91
Department of Applied Chemistry and Microbiology, Division of Nutrition and Department of Biological and Environmental Sciences, Division of Genetics University of Helsinki Minna Pekkinen REGULATION OF OSTEOBLAST DIFFERENTIATION AND GENE EXPRESSION BY PHOSPHODIESTERASES AND BIOACTIVE PEPTIDES ACADEMIC DISSERTATION To be presented, with the permission of the Faculty of Biosciences of the University of Helsinki, for public criticism in the large lecture hall, Bulevardi 18, Helsinki, on 28 March 2008, at 12 noon. HELSINKI 2008

Regulation of Osteoblast Differentiation and Gene

  • Upload
    others

  • View
    14

  • Download
    0

Embed Size (px)

Citation preview

  Department of Applied Chemistry and Microbiology, Division of Nutritionand

  Department of Biological and Environmental Sciences, Division of GeneticsUniversity of Helsinki

  Minna Pekkinen

REGULATION OF OSTEOBLAST DIFFERENTIATION AND GENE EXPRESSION BYPHOSPHODIESTERASES AND BIOACTIVE PEPTIDES

ACADEMIC DISSERTATION

To  be  presented,  with  the  permission  of  the  Faculty  of  Biosciences  of  the  University  ofHelsinki, for public criticism in the large lecture hall, Bulevardi 18, Helsinki, on 28 March2008, at 12 noon.

HELSINKI 2008

2

Supervisors

Acting Professor Christel Lamberg­AllardtDepartment of Applied Chemistry and MicrobiologyUniversity of HelsinkiFinland

Dr. Mikael AhlströmDepartment of Applied Chemistry and MicrobiologyUniversity of HelsinkiFinland

Reviewers

Acting Professor Petri LehenkariDepartment of AnatomyUniversity of OuluFinland

Adjunct Professor Anitta MahonenInstitute of BiomedicineDeparment of Medical BiochemistryUniversity of KuopioFinland

Opponent

Professor  Kalervo VäänänenInstitute of BiomedicineDepartment of AnatomyUniversity of TurkuFinland

© Minna Pekkinen 2008    Yliopistopaino

ISNB 978­952­92­3528­5 (paperback)ISNB 978­952­10­4592­9 (PDF, http//:ethesis.helsinki.fi)

3

Contents

Abstract… … … … … … … … … … … … … … … … … … … … … … … … … … … … … … 6List of original publications… … … … … … … … … … … … … … … … … … … … … … .8

Abbreviations… … … … … … … … … … … … … … … … … … … … … … … … … … … ...9

1. Introduction… … … … … … … … … … … … … … … … … … … … … … … … … .112. Review of the literature...................................................................................13

2.1 Bone biology… … … … … … … … … … … … … … … … … … … … … … … ..132.1.1    Bone tissue and bone cells.............................................................132.1.2    Osteoblastic differentiation............................................................132.1.3    Osteoblastic regulation of osteoclast differentiation......................172.1.4    Bone remodelling and its regulation...............................................182.1.5    Unbalanced bone remodelling: osteoporosis … … … … … … .........192.1.6    Glucocorticoids and bone.............................................................. .20

2.2 Cyclic nucleotide signalling and phosphodiesterases (PDEs).....................212.2.1 Cyclic adenosine monophosphate (cAMP) signalling.....................212.2.2 Cyclic guanosine monophosphate (cGMP) signalling....................222.2.3 cAMP and cGMP PDEs...................................................................23

2.2.3.1    PDE1 family............................................................................272.2.3.2     PDE2 family...........................................................................292.2.3.3     PDE3 family...........................................................................292.2.3.4     PDE4 family...........................................................................302.2.3.5     PDE5 family...........................................................................312.2.3.6     PDE6 family...........................................................................312.2.3.7     PDE7 family...........................................................................312.2.3.8    PDE8 family...........................................................................322.2.3.9    PDE9 family...........................................................................322.2.1.10  PDE10 family.........................................................................332.2.1.11  PDE11 family.........................................................................33

2.2.4 Regulation of activity and compartmentalization of PDEs.............332.2.5 Inhibitors of PDEs...........................................................................342.2.6 Bone metabolism and PDE inhibitors… … … ...… … … … ..............34

2.3 Other signal transduction pathways associated with PDEs...................... ..352.3.1 Exchange protein directly activated by cAMP (EPAC)

                             and PDEs........................................................................352.3.2 Mitogen­activated protein kinase (MAPK) and PDEs....................362.3.3 PKC and PDEs................................................................................36

4

2.4 Bioactive peptides........................................................................................362.4.1 Bioactive peptides in bone...............................................................38

3. Aims of the study… … … … … … … … … … … … … … … … … … … … … … ..… .394. Materials and methods… … … … … … … … … … … … … … … … … … … … .......40

4.1 Cell culture...................................................................................................404.1.1 Osteosarcoma cell lines.................................................................404.1.2 Human osteoblasts........................................................................404.1.3 Human mesenchymal stem cell (hMSC)­derived osteoblasts.......40

4.2 PDE inhibitors..............................................................................................414.3 Primers..........................................................................................................414.4 cAMP analyses.............................................................................................434.5 Q­sepharose chromatography.......................................................................434.6 Enzyme activity analyses.............................................................................44

4.6.1 Assay of PDE activity...................................................................444.6.2 Assay of ALP activity...................................................................44

4.7 PDE silencing...............................................................................................444.8 Isoleucine­proline­proline (IPP), valine­proline­proline (VPP) and leucine­      lysine­proline (LKP) tripeptides...................................................................454.9 Cell proliferation analyses............................................................................454.10 Mineralization analyses..............................................................................454.11 mRNA expression analyses........................................................................46

4.11.1 RNA isolation..............................................................................464.11.2 cDNA microarrays.......................................................................464.11.3 cDNA synthesis...........................................................................464.11.4 Semi­quantitative reverse transcriptase polymerase chain

                                   reaction (RT­PCR).......................................................................474.11.5 Quantitative Real­time PCR (qRT­PCR).....................................47

5. Results and discussion… ..… … … … … … … … … … … … … … … … … … … … ..485.1  Identification of PDE families in human osteoblastic cells.............485.2  Identification of PDE subtypes in human osteoblastic cells............515.3  PDE7 and PDE8A silencing by siRNAs..........................................525.4  PDE7 silencing has potential for enhancing osteogenic gene       expression and differentiation in hMSC­derived osteoblasts...........535.5  PDE7 inhibition enhances cAMP signalling, differentiation and       mineralization in hMSC­derived osteoblasts....................................565.6  Down­regulation of cAMP­PDE by dexamethasone in humanosteosarcoma cells...................................................................................585.7  Influence of dexamethasone on PDE subtype mRNA       expression in MG­63 and SaOS­2 cells............................................585.8  Effects of PDEs in osteoblasts..........................................................59

5

5.9 Influences of bioactive peptides IPP, VPP and LKP on gene       expression and proliferation in hMSC­derived osteoblasts.............60

6. Conclusions… … … … … … … … … … … … … … … … … … … … … … .................637. Acknowledgements… … … … … … … … … … … … … … … … … … … … … … … .658. References… … … … … … … … … … … … … … … … … … … … … … … … … …    66

6

Abstract

Bone  is a mineralized tissue that enables multiple mechanical and metabolic  functions to becarried out in the skeleton. Bone contains distinct cell types: osteoblasts (bone­forming cells),osteocytes (mature osteoblast that embedded in mineralized bone matrix) and the osteoclasts(bone­resorbing cells). Remodelling of bone begins early in foetal life, and once the skeletonis  fully  formed  in  young adults, almost all of  the metabolic activity  is  in this  form. Bone  isconstantly destroyed or resorbed by osteoclasts and then replaced by osteoblasts. Many bonediseases, i.e. osteoporosis, also known as bone loss, typically reflect an imbalance in skeletalturnover.

The  cyclic  adenosine  monophosphate  (cAMP)  and  the  cyclic  guanosine  monophosphate(cGMP)  are  second  messengers  involved  in  a  variety  of  cellular  responses  to  suchextracellular agents  as  hormones  and  neurotransmitters.  In  the  hormonal  regulation of  bonemetabolism,  i.e.  via  parathyroid  hormone  (PTH),  parathyroid  hormone­related  peptide(PTHrp) and prostaglandin E2 signal via cAMP. cAMP and cGMP are  formed by adenylateand guanylate cyclases and are degraded by phosphodiesterases (PDEs). PDEs determine theamplitudes of  cyclic  nucleotide­mediated hormonal  responses  and  modulate  the duration  ofthe signal. The activities of  the PDEs are regulated by multiple  inputs  from other signallingsystems and are crucial points of cross­talk between the pathways.

Food­derived  bioactive  peptides  are  reported  to  express  a  variety  of  functions  in  vivo.  Theangiotensin­converting  enzymes  (ACEs)  are  involved  in  the  regulation  of  the  specificmaturation  or  degradation  of  a  number  of  mammalian  bioactive  peptides. The  bioactivepeptides offer also a nutriceutical and a nutrigenomic aspect to bone cell biology.

The aim of this study was to investigate the influence of PDEs and bioactive peptides on theactivation  and  the  differentiation  of  human  osteoblast  cells.  The  profile  of  PDEs  in  humanosteoblast­like cells  and  the effect of glucocorticoids on  the  function of  cAMP PDEs, wereinvestigated  at  the  mRNA  and  enzyme  levels.  The  effects  of PDEs  on  bone  formation  andosteoblast  gene  expression  were  determined  with  chemical  inhibitors  and  siRNAs  (shortinterfering  RNAs).  The  influence  of  bioactive  peptides  on  osteoblast  gene  expression  andproliferation was studied at the mRNA and cellular levels.

This  work  provides  information  on  how  PDEs  are  involved  in  the  function  and  thedifferentiation  of  osteoblasts.    The  findings  illustrate  that  gene­specific  silencing  with  anRNA  interference  (RNAi)  method  is  useful  in  inhibiting,  the  gene  expression  of  specificPDEs and further, PDE7 inhibition upregulates several osteogenic genes and increases bALPactivity  and  mineralization  in  human  mesenchymal  stem  cells­derived  osteoblasts.  PDEsappear to be involved in a mechanism by which glucocorticoids affect cAMP signaling. Thismay provide a potential route in the formation of glucocorticoid­induced bone loss, involvingthe  down­regulation  of  cAMP­PDE.  PDEs  may  play  an  important  role  in  the  regulation  of

7

osteoblastic  differentiation.  Isoleucine­proline­proline  (IPP),  a  bioactive  peptide,  possessesthe potential to increase osteoblast proliferation, differentiation and signalling.

8

LIST OF ORIGINAL PUBLICATIONS

This thesis is based on the following original publications, which are referred to in the text bytheir Roman numerals (I­IV):

I Ahlström, M., Pekkinen M., Huttunen, M., Lamberg­Allardt C.  (2005): Cyclicnucleotide  phosphodiesterases  (PDEs)  in  human  osteoblastic  cells;  effect  ofPDE  inhibition on cAMP accumulation. Cellular  & Molecular Biology  Letters10(2): 305­319.

II Pekkinen, M.,  Ahlström, M.E.B, Riehle, U., Huttunen,  M.M, Lamberg­AllardtC.J.E. (2007): Effects of phosphodiesterase 7 Inhibition by RNA interference onthe  gene  expression  and  differentiation  of  human  mesenchymal  stem  cell­derived osteoblasts. Bone (accepted).

III Ahlström,  M.*,  Pekkinen,  M.*,  Huttunen,  M.,  Lamberg­Allardt,  C.  (2005):Dexamethasone  down­regulates  cAMP­phosphodiesterase  in  humanosteosarcoma cells. Biochemical Pharmacology 69: 267­275.

IV Huttunen,  M.M.*,  Pekkinen,  M.*,  Ahlström,  M.E.B,  Lamberg­Allardt  C.  J.E.(2007):  Effects  of  bioactive  peptides  isoleucine­proline­proline  (IPP),  valine­proline­proline  (VPP)  and  leucine­lysine­proline  (LKP)  on  gene  expression  ofosteoblasts, differentiated from human mesenchymal stem cells. British Journalof Nutrition 98: 780­788.

*These authors contributed equally.

These  publications  have  been  reproduced  with  the  kind  permission  of  theircopyright holders. In addition, some unpublished material is presented.

9

ABBREVIATIONSaa amino acidAC adenylyl cyclaseACEs angiotensin­coverting enzymesAKAP A­kinase anchoring proteinATF1 activating transcription factor 1ATP adenosine triphosphatebALP bone­specific alkaline phosphataseBMPs bone morphogenetic proteinsBRL 50481 3­(N,N­dimethylsulfonamido)­4­methyl­nitrobenzeneBSA bovine serum albuminBSP bone sialoproteinCaM calmodulincAMP cyclic adenosine monophosphateCdc42 cell division cycle 42cGMP cyclic guanosine monophoshateCRE cAMP response elementCREB cAMP response element­binding proteinCREM cAMP response­element modulator proteinCol1 collagen type 1Dex dexamethasonedsRNA double­stranded RNAECM extracellular matrixEHNA erythro­9­[2­hydroxy­3­nonyl] adenineEPAC exchange protein directly activated by cyclic AMPERK extracellular signal­regulated kinaseFCS foetal calf serumFGF fibroblast growth factorFSP­1 fibroblast­specific protein 1GAPDH glyceraldehyde­3­phosphate dehydrogenaseGCs glucocorticoidshMSC human mesenchymal stem cellIBMX 3­isobutyl­1­methylxanthineIGF­I insulin­like growth factor 1Ihh Indian hedgehogIPP isoleucine­proline­prolineKm Michaelis constantLKP leucine­lysine­prolineMAPK mitogen­activated protein kinaseMC mesangial cellmRNA messenger RNAM­SCF cytokine macrophage colony­stimulating factor

10

8­MMX 8­methoxymethyl­3­isobutyl­1­methylxanthineMSC bone marrow stroma mesenchymal stem cellM­SCF macrophage colony­stimulating factorMyoD transcription factor for myogenesisNO nitric oxideOCN osteocalcinOPG osteoprotegerinOPN osteopontinOsx osterixPAS per­arnt­simPCA principal component analysisPDE phosphodiesterasePGE2 prostaglandin E2 prostaglandinsPGs prostaglandinsPKA protein kinase APKB protein kinase BPKC protein kinase CPKG protein kinase GPKI protein kinase inhibitorPLC phospholipase CPPAR 2 peroxisome proliferator­activated receptor­gamma2PTH parathyroid hormonePTHR PTH/PTHrp receptorPTHrp parathyroid hormone­related peptideRANK receptor activator for nuclear factor   BRANKL receptor activator for nuclear factor   B ligandRap1 repressor activator protein 1REC                            response regulator receiverRGD                           arginine­glycine­aspartic acidRho                             ras homolog geneRNAi                          RNA interferenceRunx2                         runt related transcription factor 2siRNAs                       short interfering RNAsSox9                            SRY­box containing gene 9TGF­ transforming growth factor TRANCE receptor activator of NF­{kappa} B ligandVmax maximal velocity constantVDR 1,25­dihydroxyvitamin D3 receptorVEGF vascular endothelial growth factorVPP valine­proline­prolineWnt a combination of the genes (Wg) wingless and Int

11

1. Introduction

Bone has two main functions: it forms a rigid skeleton and plays a central role in calcium andphosphate homeostasis. Bone modelling is a process associated with growth and re­shaping ofbones  in  childhood  and  adolescence.  Throughout  human  life,  bone  tissue  is  subject  to  acontinuous turnover process whereby old bone is removed and replaced by new bone by thecoupled processes of bone resorption and bone formation. Bone is a very active organ and itsfixity after death is in sharp contrast to its ceaseless activity during life. In the adult skeleton,the  continuous  activity  largely  comprises  to  the  process  of  bone  remodelling.  Bone  isconstantly destroyed or  resorbed by  the osteoclasts  (bone­resorbing cells) and  then  replacedby the osteoblasts (bone­forming cells). This process is regulated by other bone cells, namelyosteocytes, hormones, cytokines, prostaglandins and growth factors (Raisz 1999, Ducy et al.2000, Teitelbaum 2000). A  misbalance  between  bone  resorption and  formation  may  lead  tomany  bone diseases, e.g., osteoporosis and osteopetrosis. Osteoporosis,  also known as  boneloss,  typically  reflects  an  imbalance  in  skeletal  turnover  such  that  bone  resorption  exceedsbone formation. In osteopetrosis, the misbalance is caused by a defect in osteoclastogenesis orby dysfunction of osteoclasts.

The  cyclic  nucleotides  cAMP  and  cGMP  are  second  messengers  involved  in  a  variety  ofcellular  responses  to  extracellular  agents  such  as  hormones  and  neurotransmitters.  Forexample,  parathyroid  hormone  (PTH),  an  important  hormone  acting  together  with  1,25dihydroxyvitamin D and calcitonin regulates bone metabolism, signalling via cAMP. cAMPand  cGMP  are  formed  by  adenylate  and  guanylate  cyclases  and  are  degraded  byphosphodiesterases  (PDEs).  PDEs  determine  the  amplitudes  of  cyclic  nucleotide­mediatedhormonal  responses  and  modulate  the  duration  of  the  signal.  The  activities  of  PDEs  areregulated by multiple inputs from other signalling systems and are crucial points of cross­talkbetween the pathways (Soderling and Beavo 2000, Lugnier 2006). PDEs are an important partof signal transduction in the cell, but very little is known about their role in bone cells.

During the last decade, fundamental studies have opened a new field of research dealing withbioactive  or  biogenic  substances derived  from  foods  (Meisel  2001).  The  angiotensin­converting  enzymes  (ACEs)  are  involved  in  the  regulation  of  the  specific  maturation  ordegradation  of  a  number  of  mammalian  bioactive  peptides. The  antihypertensity  effects  ofbioactive peptides on  blood  pressure  have  been  shown  in  several  studies  (e.g.  Mitsui  et  al.2004),  and  the  bioactive  peptides  offered  also  provide  a  nutrigenomic  aspect  to  bone  cellbiology.

This work was  conducted  to clarify  the  involvement of PDEs and  bioactive peptides  in  thefunction and differentiation of human osteoblast cells and to provide more insight into the roleof PDEs in the regulation of bone metabolism. Elucidation of the role of PDEs in bone cells is

12

important because bone metabolism, especially bone formation, could be affected via PDEs.The potential of bioactive peptides to increase osteogenic effects was also investigated.

13

2. Review of the literature

2.1 Bone biology2.1.1 Bone tissue and bone cells

Bone tissue  is  highly  organized  and  consists  of  a  mineral  phase  of  hydroxyapatite  andamorphous calcium phosphate crystals deposited in an organic matrix. It contains the distinctcell  types  of  osteoblasts,  osteocytes,  osteoclasts  and  lining  cells.  The  osteoblast  is  the  cellresponsible  for  collagenand  other  bone  proteins,  forming  the  bone  organic  matrix,  whichconsists of approximately 90%  type  I  collagen and 10% different non­collagenous proteins,such as osteocalcin, matrix Gla protein, bone sialoprotein, osteopontin and fibronectin (Aubinand Liu 1996). Osteoblasts also have an important role in the subsequent mineralization of thematrix. Once bone is mineralized, some of the osteoblasts are retained on the bone surface ininactive form and named as the bone­lining cells (Parfit 1994). Osteocytes are osteoblasts thathave been trapped within the bone matrix during the process of bone  formation. Osteocytescommunicate  with  one  another  and  with  cells  at  the  bone  surface  via  a  meshwork  of  cellprocesses that run through canaliculi  in the bone matrix (Kogiannis and Noble, 2007). Thus,bone cells form a functional network within which cells at all stages of bone formation, fromthe preosteoblast  to the  mature osteocyte,  remain connected. Osteoclasts are  multinucleatedcells formed by the fusion of the monocyte/macrophage family. They are responsible for boneresorption, playing a central role in formation of the skeleton and regulation of its mass. Theyattach  on  the  bone  surface,  secreting  acids  and  lysosomal  enzymes  into  the  space providedbetween their apical surfaces, and the mineralized bone surface (Ducy et al. 2000, Teitelbaum2000).

2.1.2 Osteoblastic differentiation

Bone marrow stromal mesenchymal stem cells (MSCs) have strong proliferative potential andthe capacity to form bone, cartilage, adipocytes and fibrous tissue. MSCs have been found ina  variety  of  other  adult  tissues  and  can  differentiate  into  myoblasts,  haematocytes  andpossibly  even  neural  tissues  (Reyes  et  al.  2001,  Jiang  et  al.  2002a,  2002b,  2003).  And,  inexample, osteogenic differentiation potential of human bone marrow MSCs does not decreaseby age at  late adulthood  (Leskelä et al.  2003). Commitment of MSCs  to tissue­specific celltypes  is orchestrated by morphogens, developmental  signalling pathways and transcriptionalregulators that serve as key factors (Figure 1A, Table 1).

The process of osteoblast differentiation can be subdivided in to three stages: 1) proliferation,2)  extracellular  matrix  synthesis  and  maturation  and  3)  mineralization.  Each  stage ischaracterized  by  expression  of  distinguishing  osteoblast  markers.  The  most  frequently  usedmarkers of osteoblast differentiation are alkaline phosphatase (ALP), collagen type 1 (Col1),osteopontin (OPN), bone sialoprotein (BSP), osteocalcin (OCN) and PTH/PTHrp receptor

A B

Figure 1. (A) Commitment of mesenchymal stem cells to tissue­specific cell types. Some of the known signalling proteins, growth factorsand transcription factors playing key regulator roles in the mesenchymal lineages are indicated. Also shown is noticeable plasticity betweenosteoblasts and adipocytes. (B) Osteoblastogenesis. Selected markers of the stages of maturation are used as examples. Modified fromLian et al. (2003).

TABLE 1. Some key factors in osteoblastogenesis and bone formation.Signaling protein and moleculesIndian hedgehog (Ihh)­catenin

PGE2

Functioncontrol of chondrocyte proliferation and maturation and osteoblast differentiationregulates commitment to chondrocytes (low concentration) or osteoblasts (high concentration) in Wnt signallingincreseases BMP­2 expression

Transcriptional regulators and factorsOsterixSP­3 transcription factorRunx1Runx2/Cbfa1Runx3

bone formation and osteoblast differentiationcontributes to ossification in vivopromotes chondrogenesisa master regulatory gene for activating osteoblastogenesissupports chondrocyte maturation

Growth factorsBMP2,4,7IGF­1/IGF­2Platelet­derived growth factorEpidermal growth factorEndothelinFGF

chondro­osteoprogenitor cell specificationosteoblast differentiation, a potent anti­apoptotic factor for differentiated osteoblasts and osteocytesmigration and proliferation; inhibits differentiationstimulates MSCs to differentiateincrease bone nodules in vitrosupport the development of skeletal tissues

HormonesPTH1,25(OH)2D3GlucocorticoidsSex steroids (oestrogen; androgen)Growth HormoneLeptin

numerous anabolic effects via regulation of many types of osteoblast genesa potent regulator, increasing and decreasing expression of numerous osteoblast phenotypic genespromotes osteoprogenitor cell differentiation, but increases apoptosis of mature osteoblastsstimulates osteoprogenitor proliferation and collage synthesis; increases differentiation and mineralizationstimulates proliferation and differentiationanabolic effects on osteoblast/inhibitor for bone formation

16

(PTHR).  In  general,  ALP,  BSP  and  Col1  are  early  markers  for  osteoblast  differentiation,while PTHR  and OCN appears  late, parallel with  mineralization. Osteoprogenitor  cells willarise under the appropriate stimulus from MSCs, differentiated preosteoblasts, osteoblasts andosteocytes (Figure 1B).

The  canonical  Wnt/ ­catenin  pathway  provides  early  development  cues  that  indirectlymediate the  initial cascade of gene expression  for skeletal development, bone  formation andosteoblast    differentiation  (Church  et  al.  2002,  Logan  et  al.  2004).  Bone  morphogeneticproteins [BMPs,  members  of  the  transforming  growth  factor    (TGF­ )  family] playimportant roles in directing fate decisions for MCSs. TGF­  can provide competence for earlystages  of  chondroblastic  and  osteoblastic  differentiation,  but  it  inhibits  myogenesis,adipogenesis and late­stage osteoblast differentiation.

BMPs  also  inhibit  adipogenesis  and  myogenesis,  but  they  strongly  promote  osteoblastdifferentiation (Boden et al. 1997, Hughes et al. 1995, Abe et al. 2000). The necessity of therunt  homology  domain  factor  Runx2  (Cbfa1/AML3/PEBP 1)  and  the  zing  finger  proteinosterix in bone formation has been ascertained by genetic studies (Banerjee et al. 1997, Ducyet al. 1997, Choi et al. 2000, Nakashima et al. 2002). Because Runx2 expression arises earlierthan osterix, these two factors may represent a temporal sequence of regulation crucial for theearly  and  final  stages  of  osteoblast  differentiation  (Nakashima  et  al.  2002,  Komori  2005,2006). The  influence of  transcription  factors on  lineage determination and differentiation  isrevealed  by  their  capability  to  cause  transdifferentiation  of  cells  to  other  phenotypes  whenthese  factors  are  overexpressed  or  suppressed.  For  example,  forced  overexpression  ofperoxisome  proliferation­activated  receptor  2  (PPAR 2)  in  osteoblasts  will  modify  theirphenotype  to  adipocytes,  whereas  forced  expression  of  Runx2  in  adipocytes  or  other  non­osseous  cells  results  in  acquisition  of  a  differentiated  osteoblast  phenotype  (Nuttall  et  al.1998, Lecka­Czernik et al. 1999, Jeon et al. 2003)(Figure 1).

The  developmental  expression  patterns of  transcription  factors  during  osteoblast  maturationreflect their roles as key determinants of osteoblast differentiation (Figure 1B). Many steroidand  polypeptide  hormones,  growth  factors  and  cytokines  regulate  not  only  the  growth  ofosteoprogenitors and/or their progression to mature osteoblasts but also osteoblast activity andapoptosis  (for  review,  see  Xing  and  Boyce  2005).  For  example,  PTH  stimulates  growth  ofosteoprogenitor populations, while also  inhibiting osteoblast and osteocyte apoptosis (Boyceet  al.  2002).  Leptin  is  secreted  largely  by  adipose  tissue,  but  also  by  human  mesenchymalstem  cells  (hMSCs)  and  osteoblasts  at  the  mineralization  stage.  It  enhances  proliferation,collagen synthesis, matrix deposition and mineralization (Gordeladze et al. 2002) and servesas  an  anti­apoptotic  agent  in  several  cell  systems,  including  bone  marrow  stem  cells  andosteoblasts (Shimabukuro et al. 1998, Takeda et al. 1998, Konopleva et al. 1999).

Among  important concepts  in regulation of osteoblast differentiation and activity  is  the  ideathat  local  or  endogenous  production  of  particular  growth  factors  and  cytokines  undergoes

17

changes required for osteoprogenitor differentiation to proceed and for osteoblasts to modifytheir  activity.  For  example,  insulin­like  growth  factor  1  (IGF­I)  expression  is  markedlyincreased during early osteoblast recruitment, but declines dramatically as these cells undergodifferentiation. Data from both IGF­I or IGF­I receptor­deficient mice and mice with targetedoverexpression  of  IGF­I  in  osteoblasts  show  that  IGF­I  is  involved  not  only  in  embryonicbone  development  but  also  postnatal  bone  formation  and  remodelling  (Zhao  et  al.  2000,Zhang  et  al.  2002).  Fibroblast  growth  factor  8  (FGF­8)  has  been demonstrated  to  regulatesdifferent stages of MSC differentiation toward osteogenic lineage in vitro and to increase theosteogenic capacity of bone marrow cells at the early stage of their differentiation. However,continued exposure of osteoblastic cultures to FGF­8 leads to the inhibition of bone formation(Valta et al. 2006).

2.1.3 Osteoblastic regulation of osteoclast differentiation

During  skeletal  development  and  throughout  life,  the  cells  from  the  osteoblast  lineagesynthesize  and  secrete  molecules  that  in  turn  initiate  and  control  osteoclast  differentiation(Raisz  1999).  Osteoclast  differentiation  calls  for  the  binding  of  macrophage colony­stimulating factor (M­SCF) to its receptor as well as the binding of the soluble differentiationfactor  receptor  activator  of NF­B  ligand  (RANKL)  to  its  receptor  (RANK)  on  osteoclastprecursor cells  (Teitelbaum  and  Ross  2003).  One  of  the  first  factors  cloned  that  regulatesosteoclast  differentiation  was osteoprotegerin.  Osteoprotegerin, originally  characterized  as  anovel secreted member of the tumor necrosis factor receptor superfamily, was later found toinhibit  spontaneous  or  induced bone  resorption  and  cause  osteopetrosis,  the  opposite  ofosteoporosis. Osteoprotegerin acts as a decoy receptor that binds to RANKL and prevents  itfrom interacting with its receptor (Figure 2). RANKL and osteoprotegerin, both of which areproduced by osteoblasts at different stages of maturity (Gori et al. 2000), account for some ofthe signals  in osteoblast–osteoclast communication. In conjunction with M­SCF, the RANK–RANKL–osteoprotegerin system regulates osteoclast differentiation. Thus, mice and humansdeficient  in osteoprotegerin have a high rate of bone loss (bone resorption that exceeds boneformation;  Whyte et al.  2002). As  in other high  bone  turnover  states, anti­resorptive agentscan  still  reduce  both  bone  formation  and  resorption  and  compensate for  osteoprotegerindeficiency.  The  signal(s)  that  couples resorption and  formation  remains  elusive,  althoughseveral of the anabolic ligands, such as BMPs and TGF­ß, are stored in bone matrix as bone isformed and are released during bone resorption and can thus act on osteoblasts and precursorsin  the  vicinity.  BMPs  could  also  enhance  osteoclast  differentiation  from  progenitors(Hentunen  et  al.  1998,  Itoh  et  al.  2001).  Signals  from  osteoblasts  to  osteoclasts  can  beprovided by RANKL and osteoprotegerin. In this case, preosteoblasts express a high level ofRANKL relative  to osteoprotegerin, which stimulates osteoclast differentiation and  function.More  mature  osteoblasts, by  contrast,  express  high  levels  of  osteoprotegerin  relative toRANKL, which inhibits osteoclast differentiation and function (Gori et al. 2000).

18

Figure 2. Interactions among osteoblasts and osteoclasts in bone resorption and formation andsome factors that modulate their differentiation and function.

2.1.4 Bone remodelling and its regulation

Remodelling  of  bone  begins  early  in  foetal  life,  and  once  the  skeleton  is  fully  formed  inyoung adults, almost all of the metabolic activity is in this form. The process involving boneresorption and  formation  is  initiated  by  mechanical  signals  and  is  controlled by  local  andsystemic  factors  that  regulate  osteoblast  and  osteoclast differentiation  and  function.  Boneremodelling  is  not  a  random  process and  takes  place  in  bone  remodelling units  comprisingosteoblasts, osteoclasts and  their precursors,  in which  resorption and  formation are coupled.Osteocytes are also involved in the process by soluble factors or directly through cell to cellcontacts (Zhao et al. 2002, Heino et al. 2002, 2004). Bone resorption is likely the initial eventthat occurs  in  response  to  local  mechanical  stress  signals.  The  bone  remodelling  cycleinvolves a complex series of sequential steps that are highly regulated (Raisz 1999). The mainphases  are  1)  osteoclastic  resorption,  2)  reversal  phase,  3)  preosteoblastic  migration  anddifferentiation  into  osteoblasts  4)  osteoblastic  matrix  formation  and  5)  mineralization.  Thecomplete bone  remodelling cycle at each  microscopic  site  takes about 4­6 months, with  the

19

resorption  process  lasting  about  two  weeks  and  bone  formation  several  months  (Hadjidakisand Androulakis 2006).

The  metabolic  functions of  the  skeleton  are  served  in  large  part  by  two  calcium­regulatinghormones,  PTH  and  1,25­dihydroxy  vitamin  D.  A  third  hormone,  calcitonin,  which  caninhibit bone  resorption,  is a potent  inhibitor of  bone  resorption and  is used clinically  in  thetreatment of osteoporosis. The US Food and Drug Administration has approved teriparatide,the N­terminal (1­34) fragment of recombinant human PTH, for the treatment of osteoporosis(Mishaela et al. 2002, Hodsman et al. 2005). In many countries in Europe, teriparatide is usedafter an unsuccessful course of bisphosphonate therapy or after a patient has had a previousosteoporotic fracture (Canalis 2007). PTH regulates serum ionized calcium concentration andits primary target cells  in bone tissue are  the osteoblasts (Raisz 1999). The  increase  in boneformation  after  PTH  administration  in  vivo has  been  attributed  to  increased  production  ofosteoprogenitors and  differentiation  of  osteoblasts  and  decreased  apoptosis  of  existingosteoblasts (Jilka et al. 1999). A number of studies in animals and humans have demonstratedthat  short­term, intermittent  administration  of  low  doses  of  PTH  has  a  significant anaboliceffect  on  bone  (for  review,  see  Jilka  2007).  PTH  affects  several  typical  bone  formation­associated  features  in the osteoblast such as collagen synthesis (Dempster et al. 1993), ALPactivity (Majeska et al. 1982) and synthesis of OPN (Noda and Rodan 1989). The receptor forPTH  belongs  to  the  superfamily  of  the  seven  membranes  spanning  G  protein­coupledreceptors  (Birnbaumer 1990). The  effects of PTH are known  to be  mediated by cAMP andcalcium and by the activation of protein kinase C (PKC) (Sprague et al. 1996, Ahlström andLamberg­Allardt 1997). Plasma PTH concentration tends to increase with age, and this mayproduce an increase in bone turnover and a loss of bone mass, particularly of cortical bone.

Other systemic  hormones  are  important  in  regulating skeletal growth. Growth hormone canstimulate bone formation and resorption (Rosen et al. 1998), glucocorticoids are necessary forbone cell differentiation during development (Advani et al. 1997) and thyroid hormones canalso stimulate bone resorption and formation and are critical for maintenance of normal boneremodelling  (Kawaguchi  et  al.  1994).  Probably  the  most  important  systemic  hormone  inmaintaining  normal  bone  turnover  is  oestrogen.  Oestrogen  defiency  leads  to  an  increase  inbone  remodelling  in  which  resorption  outstrips  formation  and  bone  mass  decreases.Testosterone  increases  osteoblast  formation  and  affects  bone  formation  rather  than  boneresorption  (Khosla  et  al.  2002).  Many  local  factors,  cytokines,  prostaglandins  (PGs)  andgrowth factors also regulate bone remodelling.

2.1.5 Unbalanced bone remodelling: osteoporosis

The reduced bone mass, for instance, in osteoporosis results from an imbalance between boneresorption and  formation, with  the  rate of resorption exceeding that of  formation. Excessivebone resorption causes changes  in the microstructure of the bone matrix, which make bones

20

prone  to  fracture.  Osteoporosis  has  been  proposed  to  result from  a  failure  of  resident  bonecells  to  respond  appropriately to  mechanical  load–bearing  signals  and  to  control  bone massand  architecture  (Lanyon  and  Skerry  2001).  In  the  majority  of  conditions  that lead  toosteoporosis, such as age­associated gonadal hormone deficiency, remodelling rates are high.Although  bone  formation and  resorption  are  both  increased,  the  rate  of  bone  formation isinsufficient to keep up with resorption. RANKL has been identified as a crucial cytokine forthe  formation  and  activation  of  osteoclasts.  The  effects  of  RANKL  are  physiologicallycompensated  by  the  decoy  receptor  osteoprotegerin  (OPG).  The  RANKL/OPG  system  isactivated in favour of RANKL in oestrogen deficiency, inflammation, bone malignancies andduring  treatment  with  glucocorticoids,  which  enhances  the  ratio  of  RANKL  to  OPG,  thuspromoting osteoclastogenesis, accelerating bone resorption and  inducing bone  loss (Simonetat al. 1997, Hsu et al. 1999). The balance between proliferation, differentiation and apoptosisof bone cells decides the size of osteoclast or osteoblast populations. Bone cells continuouslyreceive signals that control  their proliferation, activity and survival  from neighbouring cells,the  bone  matrix  and  hormones.  Life  history  before  and  after  menopause  and  therapeuticinterventions with drugs,  such as  sex hormones, glucocorticoids, PTH and bisphosphonates,are  therefore  partly  determined  by  the  survival  of  osteoclasts,  osteoblasts  and  osteocytes,which contribute to the bone mass and its microarchitecture (Weinstein et al. 1998, Jilka et al.1999, Borton et al. 2001, Calvi et al. 2001, Gordeladze et al. 2002, Iwata et al. 2006).

2.1.6 Glucocorticoids and bone

Glucocorticoids  (GCs)  play  an  important  role  in  modifying  the  metabolic  activity  andproliferation  of  bone  cells.  GCs  are  required  for  in  vitro  bone  nodule  formation  andmineralization, as well as for bone marrow MSC proliferation and osteogenic differentiation(Bellows et al. 1987, Shalhoub et al. 1992, Jaiswal et al. 1997, Bellows et al. 1998). Used inlong­term  therapy  as  immunosuppressive  and  anti­inflammatory  drugs,  glucocorticoids  leadto  a  reduction  in  bone  formation,  enabling  the  development  of  osteoporosis  (Eastell  et  al.1998, Patschan et al. 2001). GSs have two major inhibitory effect on bone formation: inhibitcell proliferation and decrease the population of cells capable of synthesizing type I collagen,as well as downregulate collagen gene expression in osteoblasts (Mahonen et al. 1998). GCsdecrease  osteoblast  and  osteocyte  activities  and  survival  and  have  the  opposite  effects  onosteoclasts, i.e. increased maturation and activity as well as decreased apoptosis (Weinstein etal.  1998,  O`Brien  2004,  Canalis  2005).  GCs  exposure  enhances  RANKL  expression  andinhibits  OPG  production  by  osteoblasts  (Hofbauer  et  al.  1999),  and  also  suppresses  OPGserum levels in vivo, thus elevating the RANKL/OPG ratio (Sasaki et al. 2001). In addition,GCs  enhance  the  expression  and  activity  of  PPAR 2,  a  transcription  factor  that  suppressesosteoblastogenesis and supports adipogenesis (Shi et al. 2000). Most studies have not foundan association between GCs and  increased PTH  levels (Canalis et al. 2004), but GCs mightenhance the sensitivity to PTH by changing the number of PTH receptors and their affinity toPTH (Urena et al. 1994).

21

2.2 Cyclic nucleotide signalling and phosphodiesterases (PDEs)

2.2.1 Cyclic adenosine monophosphate (cAMP) signalling

The  cAMP  is  a  second  messengers  involved  in  a  variety  of  cellular  responses  to  suchextracellular agents such as hormones, growth factors and neurotransmitters. cAMP is a partof the intracellular signalling pathway of several hormones involved in the regulation of bonemetabolism such as PTH, parathyroid hormone­related peptide (PTHrp) and prostaglandin E2(Beavo 1995, Soderling and Beavo 2000). The level of intracellular cAMP is regulated by thebalance between the activity of  two types of enzymes: adenylyl cyclase (AC) and the cyclicnucleotide  PDE.  Both enzymes  are  encoded  by  a  large  number  of genes  that  differ in  theirexpression patterns and mechanisms of regulation. The cAMP­dependent signal transductioncascades  from  cell­surface  hormone  receptors  to  AC  activation  through  the  “couplingproteins” that we know as heterotrimeric G proteins (Ross and Gilman 1977, Benovic et al.1986, Lohse et al. 1990). Nine different ACs exist in mammals. Their activities are stimulatedby  interaction  with  the    subunit  of  Gs  proteins.  Gs    subunits  are  coupled  with  differenttypes of membrane receptors in heterotrimeric complexes together with the ß and   subunits,from  which  they  dissociate  after  the  binding  of  specific  ligands  (Taussig  et  al.  1995).Degradation of cAMP is regulated by a large family of PDEs (Beavo 1995). The activities ofACs and PDEs  are  regulated positively and  negatively  by other  signalling  systems,  such ascalcium signalling [through calmodulin (CaM), CaM­dependent protein kinase II and IV, andcalcineurin],  subunits  of  other  G  proteins  (e.g.  Gi,  Go  and  Gq  proteins),  inositol  lipids  (byPKC)  and  receptor  tyrosine  kinases [through  ERK  and  protein  kinase  B  (PKB)]  (Gilman1990, Daub et al. 1996, Dessauer and Gilman 1997, Luttrell et al. 1999, Fisher et al. 2003).

Three  major  targets  of  cAMP  have  been  identified:  protein  kinase A  (PKA),  the  GTP­exchange protein  for  the  small GTPase or  the exchange protein directly activated by cAMP(EPAC)  and  the  cyclic  nucleotide­gated  ion  channels  (Kopperud  et  al.  2003).  cAMP couldalso  directly  regulate  the  mitogen­activated protein  kinase  (MAPK)  pathway  by  binding  toand activating EPAC and repressor activator protein 1 (Rap1)  (Ehses et al. 2003). Althoughthe role of Rap1 in the MAPK pathway is incompletely understood, it seems to act by bindingto and activating  mitogen­activated protein kinase kinase kinase B­Raf  and/or  inhibiting theRas­Raf (Ras as a G protein and Raf as a mitogen­activated protein kinase kinase) pathway(Zhong 1995, Stork and Schmitt 2002).

PKA is composed of a complex of two regulatory (R) subunits and two catalytic (C) subunits.Four  genes  (RI ,  RIß,  RII   and  RIIß)  encode  the  R  subunits,  and  three  (C ,  Cß,  C   andPrKX) encode the C subunits (Skalhegg and Tasken 2000). PKA is activated by the binding ofcAMP to the R subunits, which induces their dissociation from the C subunit (for review, seeTaylor  et  al.  2005)  Apart  from  the  R  subunits,  PKA  activity  is  down­regulated  by protein

22

kinase  inhibitors (PKIs). Three known isoforms of PKI , ß and  ) are encoded by differentgenes,  each  of  which has  a  specific  expression  pattern.  Interestingly,  PKI  could act  as  achaperone  for  nuclear  export  of  the  C  subunit,  thereby  interfering  with  the  extent  of  PKAactivity  in  the  nucleus  (Taylor  et  al.  2005,  Dalton  and  Dewey  2006).  An  important stepforward  in  our  understanding  of  the  action  of  PKA  has  been the  identification  of  PKA­anchoring proteins (AKAPs). These proteins allow specificity in cAMP signal transduction byplacing PKA close to specific effectors and substrates (Huang et al. 1997).

A  large number of proteins have  been  identified as substrates for PKA. PKA  modulates  thecAMP  response  by  phosphorylating  other  components of  the  cAMP  pathway  such  asreceptors,  ACs  and  PDEs.  Depending  on  the  cellular  system  involved,  phosphorylation ofthese proteins  results  in an enhancement of  the cAMP signal  (e.g.  repressing the activity ofPDE1) (Giorgi et al. 2002, Goraya et al. 2004) or in a rapid down­regulation of the pathway toobtain  a  transient  effect  (e.g.  by  repressing  the  activity  of  AC  V  and  AC  VI,  or  the  2adrenergic receptor) (for review, see Beazly and Watts 2006). Regulation of transcription byPKA  is  mainly  achieved by  direct  phosphorylation of  transcription  factors.  cAMP responseelement­binding  protein  (CREB),  cAMP  response­element  modulator  protein  (CREM)  andactivating transcription factor 1 (ATF1), originally identified as the activators that respond tocAMP,  are  phosphorylated  by  PKA  in  their  activation  domains. Several  lines  of  evidencesupport  the  notion  that  CREB, CREM  and  ATF1  can  be  phosphorylated  by  many  differentkinases  (Ziff 1990,  Foulkes et al.  1991, Rehfuss  et al.  1991,  Lalli  and Sassone­Corsi 1994,Sassone­Corsi 1995).

2.2.2 Cyclic guanosine monophosphate (cGMP) signalling

Natriuretic peptides and nitric oxide (NO) activate the cGMP/cGMP­dependent protein kinaseG  (PKG)  signalling  pathway.  The  role  of  cGMP  as  a  second  messenger  has  long  beenrealized, but the mechanism of action has only been addressed  in the  last couple of decades(Francis  and  Corbin  1994,  Forte  and  Currie  1995).  Most  of  cGMP's  downstream  actions(Eigenthaler et al. 1999), and possibly  some of cAMP's actions (White et al. 2000),  involveactivation of PKGs. Unlike cAMP, however, cGMP has been found to regulate other proteinclasses such as ion channels (Mery et al. 1991, Delay et al. 1997, Hagen et al. 1998, Anthonyet al. 2000) and PDEs (Francis et al. 1994). The two forms of mammalian PKG have distinctand overlapping tissue distributions. Type I  isoforms exist as alpha and beta splice variants,are  cytosolic  and  have  been  recognized  in  the  brain  (Tsou  et  al.  1993),  smooth  muscle(Komalavilas and Lincoln 1996), platelets, cardiac myocytes (Kwak et al. 1995) and kidney(Gambaryan  et  al.  1996).  The  type  II  isoform  is  myristylated  and  membrane­bound  and  ispresent  in intestinal mucosa (French et al. 1995), kidney (Gambaryan et al. 1996), brain (el­Husseini  et  al.  1995)  and  bone  (Pfeifer  et  al.  1996).  Both  PKGs  contain  amino­terminalleucine zipper motifs, and, in contrast to PKA, exist as homodimers in native tissues (Francisand Corbin 1994).

23

Little  is  known  about  PKG  regulation  of  gene  expression,  and  gene  expression  profiling  isjust beginning to contribute to the growing  list of cGMP­regulated genes (Pilz and Broderik2005). Expression of the immediate early genes c­fos and junB, but not of c­jun (a componentof  the  transcription  factor  activator  protein  1)  or  proto­oncogene  junD,  is  increased  uponactivation  of  the  cGMP  pathway.  Furthermore,  this  induction  can  be  inhibited  with  theselective  PKG  inhibitor  KT5823  (Haby  et  al.  1994).  cAMP  response  element  (CRE)dependent gene  transcription  is  induced  by  transfected PKG, but weakly,  as compared withPKA (Collins and Uhler 1999). Both nitric oxide and atrial natriuretic peptide activate the c­fos promoter through a guanylyl cyclase­ and PKG­dependent mechanism (Gudi et al. 1996,Idriss  et  al.  1999,  Lincoln  et  al.  2006).  Many  other  genes  probably  are  targeted  bycGMP/PKG,  including vascular  endothelium growth  factor  (VEGF) and  the VEGF receptor(Tuder et al. 1995, Dulak et al. 2000), smooth muscle actin and calponin (Boerth et al. 1997,Lincoln  et  al.  1998),  but  whether  these  genes  are  activated  directly  by  PKG  or  indirectlythrough cross­talk with other signalling systems is obscure. Re­expression of PKG in culturesof mammalian cells that produce low levels of endogenous protein has been helpful in testingthe requirement for PKG in various signal transduction systems. Gudi et al. (1997) observedthat a specific cGMP­dependent nuclear localization signal controls PKG translocation to thenucleus  and  transactivation  of  the fos  promoter.  These  data  suggest  that  the  cGMP/PKGsystem  plays  a  potentially  important  and  direct  role  in  transcriptional  regulation.  However,Collins  and  Uhler  (1999)  observed  exclusively  cytoplasmic  localization  of  both  PKGisoforms and could only demonstrate nuclear translocation and transcriptional activation whenthe protein was  mutated to a chimeric  form containing  the C  subunit of PKG and  the PKAcatalytic domain. These findings indicate that PKGs, in contrast to PKA, may, in fact, utilizeadditional cytoplasmic factors to regulate gene expression.

2.2.3. cAMP and cGMP PDEs

The  second  messengers,  cAMP  and  cGMP,  are  inactivated  by  the  cyclic  nucleotide  PDEs(Figure  3).  Mammalian  PDEs  have  an  HD  domain  (a  superfamily  of  metal­dependentphosphohydrolases;  Aravind  and  Koonin  1998),  the C­terminal  half  and  show  high  affinityfor cAMP and/or cGMP. Protein domains  involved  in regulation of PDE enzymatic activityand  subcellular  localization  are  mainly present  in  the  N­terminal  half.  Some  PDEs  havephosphorylation sites targeted by protein kinases and lipid modification sites. Approximately270  aa  in  the  C­terminal  catalytic domain are  conserved,  with a  sequence  identity  of  35%­50%  among  different  PDE  families  (Figure  4). Some  PDE  families  are  composed  of  2­  4subfamily genes  showing sequence  identity  of  more  than  70%  and  having  identical  proteindomain organization. Twenty­one PDE genes have been identified in humans, rats, and mice.They are categorized into 11 structurally, biochemically and pharmacologically distinct PDEfamilies, PDE1­PDE11 (Manganiello et al. 1995, Soderling et al. 1998a, 1998b, Fawcett et al.

24

Figure  3.  Signalling  using  3`5`  cAMP  and  cGMP. A.  Natriuretic  peptide  receptor  (NPR)activates guanylate cyclase (GC) upon receptor binding, followed by cGMP production fromprecursor GTP. Soluble GC, activated by NO also catalyses the production of cGMP. Threeknown cGMP target molecules, protein kinase G (PKG), cyclic nucleotide phosphodiesterase(PDE)  and  cyclic  nucleotide  gated  channels  (CNG  channel),  may  be  activated,  leading  toinitiation  of  cellular  responses. B.  Receptors  (GPCRs)  bind  specific  receptor  proteins.Receptors  then  interact with G proteins  found at  the cytosol  face of  the  membrane,  causingthem  to  bind  GTP,  change  conformation  and  subsequently  activate  the  enzyme  adenylylcyclase (AC). This enzyme produces cyclic AMP from adenosine triphosphate (ATP) withinthe  cell  interior.  Cyclic  AMP  then  exerts  its  action  in  the  cell  by  binding  to  and  activatingforms  of  protein  kinase  A  (PKA).  This  family  of  enzymes  elicits  the  phosphorylation  ofspecific targets proteins, causing functional changes within the cell. Modified from Omari etal. (2007).

25

Figure  4.  General  structure  of  phosphodiesterases.  The  different  domains  of  the  PDEs  aredrawn  as  boxes  connected  by  wires.  PDEs  share  a  common  structure,  with  a  conservedcatalytic domain  (light blue)  flanked by  regulator domains (grey and white)  the N­terminusand  C­terminus.  The  white  box  corresponds  to  regulator  domain  1,  where  severalphosphorylation sites or allosteric binding sites have been mapped. Modified from Mehats etal. (2002).

2000,  Conti  and  Beavo  2007),  some of  which  specifically  hydrolyse cAMP  (PDE3,  PDE4,PDE7 and PDE8) and others cGMP (PDE5, PDE6 and PDE9). The PDEs  belonging  to thePDE1,  PDE2,  PDE10  and  PDE11  families  are  dual­substrate  enzymes,  hydrolysing  bothcAMP and cGMP (Soderling et al. 1998a, 1998b, Fawcett  et  al. 2000). Most PDE familiescontain multiple genes (subtypes), and most of these genes can produce multiple transcripts orvariants (Conti and Jin 2000). The unique characteristics of each PDE gene family are definedby  protein  domains  located  in  the  N­terminal  to  the  catalytic unit.  As  shown  in  Figure  5,approximately half of the PDE gene families (PDE2, PDE5, PDE6, PDE10 and PDE11) havea protein domain termed GAF in tandem and are therefore designated a GAF­PDE

26

Figure 5. Domain and motif organization of each PDE family. Modified from Soderling andBeavo (2000).

27

subfamily.The  known  functions  of  GAF  domains are  cGMP  binding­mediated  allostericregulation  and  dimerization of  GAF­PDEs.  Some  GAF  domains  have  also  been  reported  tobind cAMP. Other PDEs (PDE1, PDE3, PDE4 and PDE7–9) have no GAF domain and belongto  the  non–GAF­PDE  subfamily.  PDE1 contains  a  Ca2+/CaM­binding  site,  PDE3  has  atransmembrane domain,  PDE4  has  upstream  conserved  regions  (UCRs)  and  PDE8 has  aresponse  regulator  receiver  (REC)  domain  and  a  per–arnt–sim  (PAS) domain.  PDE7  andPDE9 have no specific protein domain in addition to the PDE catalytic domain.

Each  PDE  family  is  distinguished  functionally  by  its  unique  combination  of  enzymaticcharacteristics and pharmacological inhibitory profiles (Table 2). Different families, subtypesand variants are varied in different tissues and cell types. One of the main roles of the PDEs isto  determine  the  amplitude  of  cyclic  nucleotide­mediated  hormonal  responses  and  tomodulate  the duration of  the  signal. Therefore, the PDEs can  function as  modulators of  thecyclic nucleotide­mediated signal transduction pathways (Manganiello et al. 1995, Mehats etal.  2002,  Housley  et  al.  2003,  Conti  et  al.  2003).    The  activities  of  PDE  are  regulated  bymultiple  inputs  from  other  signalling  systems  and  PDEs  are  crucial  in  cross­talk  betweenpathways.

In  the  nomenclature  used  for  each  PDE  isozyme  (e.g. HsPDE1A1),  the  first  two  lettersindicate  the animal  species and the  first Arabic number  after PDE designates  the PDE genefamily. This number is followed by a single capital letter indicating a distinct subfamily gene.The  last  Arabic  number  indicates a  specific  splice  variant  or  a  specific  transcript  generatedfrom a unique transcription initiation site (http://depts.washington.edu/pde/pde.html).

2.2.3.1 PDE1 family

Three subfamily genes, PDE1A–C with various splice variants encode Ca2+/CaM­dependentcAMP­  and  cGMP­hydrolysing  PDEs.  PDE1  is  mainly  present  in  the  cytosolic  fraction;however  PDE1  has  also  been  found  in  the  fibres  of  several  neurons  from  the  dorsal  rootganglion (Giorgio et al. 2002). In humans, PDE1A shows high affinity for cGMP (Loughneyet al.  1996). PDE1B hydrolyses cGMP with a Km  value  lower and a Vmax value higher  thanthose  for  cAMP  (Bender  et  al.  2005).  High  affinity for  both  cAMP  and  cGMP  is  observedwith PDE1C (Loughney et al. 1996). In cells expressing PDE1, hormones that raise cytosolicCa2+ would  active  PDE1,  in  this  way  diminishing  the  cyclic  AMP  response  to  hormonesstimulating  cAMP  or  cGMP  synthesis.  Sustained  Ca2+ entry  into  the  cells  is  required  toactivate PDE1A in the astroma cell line, and PDE1A reportedly cannot discriminate betweenthe different sources of Ca2+ entry (Goraya et al. 2004). By contrast, in vivo phosphorylationof PDE1 would likely result in the potention of cAMP or cGMP accumulation, involving theelevation of cytosolic Ca2+ and activation of guanylyl or adenylyl cyclases, and would play amajor  role  in  the  amplification  and  prolongation  of  cyclic  nucleotide  effects.  Differentalterations  in  tissue  and  cell  PDE1  isozymes  may  occur  by  short­term  (phosphorylation,proteolysis) or long­term regulation (at the mRNA level) in response to hormonal stimulation

28

TABLE 2. Superfamily of phosphodiesterases.

Family  Substrate Property Numberof genes

Selective inhibitors  IBMXSensitivity

PDE1 cAMP,cGMP

Ca2+­CaM­stimulated

3 8­methoxy­IBMX,phenothiazines,nimodipine

     +

PDE2  cAMP,cGMP

cGMP­stimulated  1 erythro­9­[2­hydroxy­3­nonyl]adenine (EHNA)

     +

PDE3  cAMP,cGMP

cGMP­inhibited  2 Cilostamide,milrinone

     +

PDE4  cAMP cGMP­insensitive

4 Rolipram,roflumilast, ArifloTM

     +

PDE5  cGMP PKA/PKG­phosphorylated

1 Zaprinast, sildenafil(Viagra TM)

     +

PDE6  cGMP Transducin­activated

4 Zaprinast, DMPPO,E4021, Sildenafil

     +

PDE7  cAMP Rolipram­insensitive

2 BRL 50481, ICI242       +

PDE8  cAMP Rolipram­insensitive

2 Unknown      ­

PDE9  cGMP cGMP highaffinity

1 Unknown      ­

PDE10  cAMP,cGMP

Unknown 1 Papaverine      +

PDE11  cAMP,cGMP

Unknown 1 Unknown      +

and probably play a role in spatial­temporal regulation of PDE1 activity (Spence et al. 1997,Michibata  et  al.  2001,  Yan  et  al.  2001).  In  humans,  PDE1A5  and  PDE1A6  expression  isbrain­specific,  whereas  PDE1A1  and  PDE1A4  expression  is  omnipresent,  but  particularlyhigh  in  the  kidney,  liver,  pancreas  and  thyroid  gland  (Michibata  et  al.  2001,  Fidock  et  al.2002). PDE1A10 expression is testis­specific (Michibata et al. 2001).

29

2.2.3.2 PDE2 family

PDE2A hydrolyses both cGMP and cAMP with similar maximal rates and relatively high Km

values. PDE2A is allosterically stimulated by cGMP binding to its GAF domain (Martinez etal.  2002),  enabling  the  simultaneous  regulation of  both  cAMP  and  cGMP  signalling.  PDE2was shown to play a major feedback role by restoring the basal level in cyclic nucleotides inresponse  to  hormonal  stimulation  of  the  adrenal  gland  (MacFarland  et  al.  1991).  Threevariants  of  a  single  gene  were  cloned  for  PDE2:  PDE2A1,  PDE2A2  and  PDE2A3(Sonnenburg et al. 1991, Yang et al. 1994, Rosman et al. 1997). PDE2 activity is up­regulatedin vivo at the post­transcriptional level by PKC under 4 beta­phorhol 12­myristate 13 acetate(Geoffroy et al. 1999). In endothelial cells, PDE2  is up­regulated during phenotype changes(Keravis  et  al.  2000)  as  well  as  under  stimulation  by  VEGF  (Favot  et  al.  2004),  indicatingPDE2  participation  in  endothelial  cell  proliferation.  PDE2  protein  is  mainly  present  in  theadrenal medulla, heart, rat ventricle (Yanaka et al. 2003), brown adipose tissue (Coudray et al.1999), liver and brain.

2.2.3.3 PDE3 family

PDE3A  and  PDE3B,  subfamily  genes  of  PDE3,  show  a  high  affinity  for  both  cAMP  andcGMP. A low Vmax value for cGMP compared with that for cAMP allows cGMP to functionas a competitive inhibitor for cAMP hydrolysis (Degerman et al. 1997). Therefore, PDE3s aretermed cGMP­inhibited cAMP PDEs. The presence of a 44­aa insert in the catalytic domain isa  unique  characteristic  of  the  PDE3 family.  Another  special  feature  is  the  presence  of  N­terminal hydrophobic  membrane  association  domains  (Wechsler  et  al.  2002).  Short­termactivation  of  PDE3  was  firstly  demonstrated  in  rat  fat  cells  in  response  to  insulin  andisoprenaline  stimulations,  which  induce  PDE3  phosphorylation  (Degerman  et  al.  1990).  Asingle phosphorylated serine site (Ser 302) was identified (Rahn et al. 1996). This site couldbe phosphorylated either by PKA or by PKB, but  the major site of PKB phosphorylation  isSer 273 (Kitamura et al. 1999). Possible activation of PDE3B by phosphoinositide 3­kinasephosphorylation (Rondinone et al. 2000) was recently  implicated  in the hypothalamic actionof leptin on feeding (Zhao et al. 2002) as well as in  ­cell insulin secretion (Zhao et al. 1998).Furthermore, hormonal  stimulation  (insulin, glucagon) was shown  to activate  in  vivo PDE3associated  with  the  Golgi­endosomal  fraction  (Geoffroy  et  al.  2001).  PDE  3A  is  mainlypresent  in  the  heart  muscle  cells,  platelets,  vascular  smooth  muscle  cells  and  oocytes(Reinhardt et al. 1995, Degerman et al.1997, Wechsler et al. 2002), whereas PDE3B is mainlyassociated with adipocytes, hepatocytes and spermatocytes (Miki et al. 1996).

30

2.2.3.4 PDE4 family

The PDE4  family exclusively hydrolyses cAMP (Km=2­4 µM). Currently,  the PDE4  familyrepresents the largest and most studied PDE family, comprising four highly similar subfamilygenes,  PDE4A­D,  that  encode  rolipram­sensitive  PDEs.  The PDE4  family  includes  variousalternative  splice  variant  groups  encoding  long  PDE4  and  short  PDE4  isozymes,  with  aminimum of 35 different PDE4 proteins (Houslay 2001) based on the presence or absence ofN­terminal  upstream  conserved  region  (UCR)  domains  (Bolger  et  al.  1993). The  long­formvariants  have  UCR1,  linker  region  (LR)  1,  UCR2, LR2,  and  a  catalytic  domain.  The  short­form variants contain LR1­UCR2­LR2 and UCR2 (truncated)­LR2 in the N­terminal region.UCR1, which includes one PKA phosphorylation site, is connected to UCR2 by LR1. UCR2has a hydrophilic N­terminal region, which intramolecularly interacts with the hydrophobic C­terminal  portion  of  UCR1  (Beard  et  al.  2000).  UCR1  and  UCR2  are  involved  in  PDE4enzymatic regulation (Sette and Conti 1996, Hoffman et al. 1999, Grange et al. 2000) throughUCR2 interaction with the catalytic domain (Lim et al. 1999) and have also been reported toparticipate in PDE4 dimerization (Richter and Conti 2002). Although the PDE4 catalytic sitebinds  to  the  competitive  inhibitor  rolipram, the  affinity  of  this  site  for  rolipram  can  changemarkedly  depending on  the  conformation  of  the  enzyme.  PDE4  activity  is  controlled  byphosphorylation,  a  relationship with a protein or  endogenous mediator  and proteolysis. Thepresence of an acceptor site in UCR1 for PKA­mediated phosphorylation allows a fast changein  PDE4  activity.  In  vivo,  this  augmentation  of  PDE4  activity  was  shown  as  a  result  of  aextended increase of cAMP resulting from hormonal stimulation; this was demonstrated to bea short­term feedback mechanism allowing cAMP  levels  to return to the basal cellular  state(Sette  et  al.  1994,  Oki  et  al.  2000).  An  extended  accretion  of  cAMP  induces  a  long­termregulation  of  PDE4  induction.  Follicle­stimulating  hormone  and  dibutyryl  cAMP  treatmentwas shown to start the specific induction of  short­form PDE4D1 and PDE4D2 in rat Sertollicells,  whereas  the  expression  of  the  long­form  PDE4D3  was  unchanged  (Swinnen  et  al.1991).  Conti  et  al.  (1995)  suggested  that  the  PDE4D  gene  had  two  distinct  transcriptionalunits,  the  unit  controlling  the  short­form  expression  being  up­regulated  by  cAMP.Correspondingly,  cAMP­dependent  up­regulation  was  shown  for  PDE4A  and  PDE4B  inU937  (Torphy  et  al.  1995),  for  PDE4B2  in  human  monocytes  (Manning  et  al.  1996)  and,more  recently,  in  human  myometrial  cells  (Oger  et  al.  2002).  Furthermore,  the  down­regulation of PDE4A and PDE4B could be connected to PDE4D up­regulation, as shown  inendothelial cells during phenotypic variance (Keravis et al. 2000), whereas PDE4A was up­regulated  and  PDE4B  and  PDE4D  were  down­regulated  in  the  brain  during  fluoxetinetreatment  (Miro  et  al.  2002).  In  humans,  PDE4A­PDE4D  expression  is  fundamentallyubiquitous (Engels et al. 1994, 1995, Bolger et al. 1997, Wang et al. 1999, Wang et al. 2003),with  a  variant­specific  tissue  distribution  pattern,  but  PDE4A  has  been  reported  to  berelatively  high  in  the  brain  (Bolger  et  al.  1993)  and  PDE4B  low  in  the  lung  and  absent  inblood (Engels et al. 1995).

31

2.2.3.5 PDE5 family

To  date,  only  one  gene  product  has  been  reported  for  this  family,  PDE5A,  which  has  twoGAF domains (GAF A and GAF B) in the N­terminal half and specifically hydrolyses cGMP.The  GAF  A  domain  of  PDE5A  has  been reported  to  be  responsible  for  this  enzyme`sallosteric binding to cGMP (Liu et al. 2002), and therefore PDE5A is termed a cGMP­bindingcGMP­specific PDE. One PKG­ and PKA­dependent phosphorylation site  in  the N­terminalregion  is related to activation of  the PDE5A enzyme (Corbin et al. 2000). cGMP binding tothe PDE5A GAF A domain promotes phosphorylation, which not only activates the catalyticfunction  but  also  increases cGMP­binding  affinity  (Corbin  et  al.  2000,  Francis  et  al.  2002,Zoraghi et al. 2005). In humans, PDE5A transcripts are abundantly found in various tissues,especially  in  smooth  muscle  tissues,  and  are  also  detected  in  platelets  (Yanaga  et  al.  1998,Kotera  et  al.  1999).  PDE5A1 and  PDE5A2  transcripts  are  widely  distributed  (Kotera  et  al.1999, Lin et al. 2000). In contrast, specific expression of PDE5A3 in  smooth and/or cardiacmuscle has been suggested (Lin et al. 2000).

2.2.3.6 PDE6 family

In  retinal  rod  and  cone  cells,  the  level  of  cGMP,  a  second  messenger in  visual  signaltransduction, is tightly controlled through regulated cGMP hydrolysis by PDE6A­C subfamilygenes (photoreceptor PDEs) (Cote 2004). Light­activated transducin stimulates PDE6 activityby  removing  the  inhibitory  subunit  .  Membrane  hyperpolarization is  caused  by  cGMPelimination, leading to an electrical cellular response. PDE6  and PDE6ß subunits encoded bythe PDE6A and PDE6B genes, respectively, form a holoenzyme, rod PDE, with two copies ofthe smaller inhibitory subunit   (PDE6 ) encoded by PDE6G. The GAF A domain is relatedto  PDE6 ß  heterodimerization  (Muradov  et  al.  2003). In  cone  cells,  a  homodimer  of  2  'subunits  (PDE6 ')  encoded  by PDE6C  comprises  cone  PDE  with  cone­specific  inhibitorysubunits encoded by PDE6H. Thus, regulation of PDE6 activity by small inhibitory subunits isa unique aspect of  this PDE family. Several different genetic diseases   caused by defects  inPDE6  function  or  expression  have  been  characterized;  in  humans,  for  instance,  autosomaldominant congenital  stationary  night blindness and autosomal  recessive  retinitis pigmentosahave been attributed to PDE6B mutations (Gal et al. 1994).

2.2.3.7 PDE7 family

This  family  includes two genes, PDE7A and PDE7B, that encode rolipram­insensitive high­affinity  cAMP­specific  PDEs  (Km  value  approximately  0.2 µmol/L)  (Michaeli  et  al.  1993).The  PDE7  family  contains  neither  GAF  domains  nor  regulatory  domains  (Beavo  1995).PDE7A  and  PDE7B  subtypes  show  approximately  70%  homology  (Gardner  et  al.  2000,Sasaki et al. 2000). A PKA pseudosubstrate site  is present  in the N­terminus of  the PDE7Asubfamily.  Alternative  splicing  for  PDE7A  and  tissue­specific  expression  of    PDE7  splice

32

variants have been identified (Bloom and Beavo 1996, Han et al. 1997). PDE7A1 mRNA andprotein are distributed ubiquitously across human proinflammatory and immune cells (Smithet al. 2003). On the other hand, PDE7A2, which is generated by 5'­splicing and differs fromPDE7A1 at its N­terminus (Bloom and Beavo 1996, Han et al. 1997), has never been detectedat the protein  level, despite unequivocal  identification of  its mRNA (Smith et al. 2003). Thedistribution of PDE7A3 is largely unknown, but in human it has been found in T­lymphocytes(Glavas et  al.  2001)  and  may  also  be  present  in  many  PDE7A1­expressing cells,  as  bothtranscripts are probably regulated by the same promoter (Torras­Llort and Azorin, 2003). Inrats, three N­terminal splice variants, PDE7B1­3, exist. Only PDE7B1 has been identified inhumans  and  mice  and  it is  abundant  in  the  brain,  liver,  heart,  thyroid  glands  and skeletalmuscles (Gardner et al. 2000).

2.2.3.8 PDE8 family

PDE8A  and  PDE8B  are  the  subfamily  genes  of  PDE8.  PDE8s  are  high­affinity cAMP­specific PDEs  insensitive  to  rolipram and 3­isobutyl­1­methylxanthine (IBMX) (Gamanumaet  al.  2003),  and  contain  response  regulator  receiver  (REC)  and  PAS  domains  in  the  N­terminal portion.  The  REC  domain  functions  as  a  receiver  of  signals  from the  sensorcomponent in a 2­component signal transduction system in lower organisms. The PAS domainis  involved  in  the  binding  of small  ligands  and  protein–protein  interactions  (Huang  et  al.1993).  However, regulation  of  PDE8s  via  REC  or  PAS  domain  is  unknown,  and  obviousPDE8  activity  has  not  yet  been  demonstrated  in  either tissue  or  cell  extracts.  PDE8Atranscripts  are  expressed  in  various  tissues  and  are  abundant  in  the  testis,  ovary,  smallintestine and colon (Fisher et al. 1998). The expression of PDE8B seems to be confined to thethyroid gland and the brain (Kobayashi et al. 2003).

2.2.3.9 PDE9 family

PDE9A is the only reported subfamily gene of PDE9, although more than 20 splice variantshave  been observed  in  humans,  indicating  that  the PDE9A gene appears  to have a complexregulation  of  expression  (Rentero  et  al.  2003).  PDE9A  specifically  hydrolyses  cGMP  withhigh affinity (Fisher et al. 1998, Guipponi et al. 1998). However, no reports are available onthe regulation of PDE9A activity or on the presence of endogenous PDE9A activity in eithertissue or cell extracts. IBMX­insensitive cGMP PDE activity, which has been shown to date,is  likely  attributable  to  PDE9.  PDE9  mRNA  is  highly  conserved  between  species  and  iswidely distributed throughout the rodent brain (Van Staveren et al. 2002). PDE9A mRNA isexpressed in the spleen, small intestine, brain, colon, prostate, kidney and placenta (Fisher etal. 1998, Guipponi et al. 1998). In humans, the PDE9A5 splice variant is highly expressed inimmune tissues (Wang et al. 2003).

33

2.2.3.10 PDE10 family

PDE10A contains two N­terminal GAF domains and hydrolyses both cAMP and cGMP. Twomajor variants, PDE10A1 and PDE10A2, and several minor variants have been identified inhumans  (Fujishige  et  al.  1999a,  2000).  High  affinity  for  cAMP  inhibits  cGMP  hydrolysis,making this enzyme a cAMP­inhibited dual­substrate PDE. Among newly discovered PDEs,the enzymatic activity of PDE10A is clearly demonstrated in tissue extracts (Fujishige et al.1999b).  The  enzymatic  activity  of a  chimeric  construct  of  the  PDE10A  GAF  domain  andcyanobacterial adenyl  cyclase  is  stimulated  by  cAMP,  suggesting  a  possible allostericmodulation of PDE10A activity by cAMP (Gross­Langenhoff et al. 2006). PDE10 transcriptsare  particularly  abundant  in  the  brain,  thyroid  and  testis.  The  PDE10  family  was  recentlyshown  to  be  associated  with  a  progressive  neurodegenerative  disease,  Huntington`s  disease(Hebb et al. 2004, Hu et al. 2004).

2.2.3.11 PDE11 family

The  most  recently  identified  family,  PDE11,  contains  four  N­terminal  variants,  PDE11A1­PDE11A4. A full­length form, PDE11A4, contains two GAF domains and a catalytic domain.PDE11A  hydrolyses  both  cAMP  and  cGMP  with  similar Km values  (Fawcett  et  al.  2000,Hetman et al. 2000, Yuasa et al. 2000). Although a cyanobacterial adenyl cyclase fused withPDE11A4  GAF  domains  is  activated  by  cGMP  (Gross­Langenhoff  et  al.  2006),  no  reportsexist on the allosteric regulation of  the PDE11A enzyme. PDE11A activity has not yet beenclearly demonstrated in tissue or cell extracts. In humans, PDE11A expression is strong in theprostate and moderate in the testis and several other tissues (Fawcett et al. 2000, Yuasa et al.2000).

2.2.4 Regulation of activity and compartmentalization of PDEs

PDEs  exist  universally,  from prokaryotes  to  eukaryotes.  In  all  organisms,  multiple  PDEsappear  to  be  involved  in  tight  feedback  regulation  of  cAMP  and  cGMP.  Observations  alsoimply that different combinations of PDEs can regulate the same general function in differentspecies.  Divergent  PDE  expression  patterns  have  been  detected  in  the  hearts  of  the  rodentsand humans. Two general types of PDE regulation occur. Short­term regulation involves theactivation  of  second  messenger  pathways  and  consequent  allosteric  modulation  of  enzymeactivity, and  long­term regulation occurs  through  increased enzyme synthesis. Several PDEfamilies  are  subject  to  one  or  both  forms  of  regulation.  Functional  evidence  indicates  thatcyclic nucleotide signals are confined in compartments and that different PDEs, as well as thepresence  of  PDE/PKA  or  PDE/PKG  complexes,  may  contribute  to  establishment  orstabilization  of  these  compartments.  However,  the  exact  mechanism  restricting  cyclic

34

nucleotide  diffusion  remains  unclear,  and  convincing  evidence  for  PDE­generated  cyclicnucleotide gradients is still lacking (for review, see Conti and Beavo 2007).

2.2.5 Inhibitors of PDEs

The PDEs are targets for a number of synthetic PDE isozyme inhibitors. Theophylline was thefirst chemical inhibitor described for PDEs (Butcher and Sutherland 1962). One decade later,a  new  xanthine  analogue,  IBMX,  was  shown  to be 100­fold  more  potent  than  theophylline(for  review, see Chasin and Harris 1976). These two compounds are both non­specific PDEinhibitors. The last decade, numerous more or less selective inhibitors have been designed andsynthesized  for scientific and pharmacological purposes. Some of them  inhibit PDE activityvery  potently  in  the  nanomolar  and  subnanomolar  ranges  and  are  specific  to  certain  PDEfamilies. Thus, in the span of several decades, the potency of PDE inhibitors has increased by108,  and  the  specificity  of  some  of  them  by  104.  Recently,  several  selective  inhibitors  forindividual  genes  within  the  PDE  family  have  been  discovered  and  synthesized,  e.g.  OPC­33540  inhibits  PDE3A  (IC50=  0.32  nM)  (Sudo  et  al.  2000)  and  IC242  inhibits  PDE7Aselectively, with an  IC50 of 0.37 µM (Lee et al.  2002a, 2002b). Efforts are now  focused ondeveloping  third­generation  inhibitors  that  are  subtype­specific.  PDE  inhibitors  areexperimental  tools  singularly  useful  in  studies  of  the  role  of  PDEs  in  cyclic  nucleotidesignalling  and  also  represent  a  novel  class  of  pharmacological  agents  suited  for  “signaltransduction  pharmacotherapy”  (Levitzki  1997).  Currently,  PDE1,  PDE2,  PDE3,  PDE4,PDE5,  PDE6,  PDE7  and  PDE10  have  more  or  less  selective  inhibitors,  but  no  selectivechemical inhibitors have yet been described for PDE8, PDE9 or PDE11 (Table 2). Because oftheir great market potential and therapeutic relevance, PDE  inhibitors became recognized asimportant  therapeutical  agents  for  a  variety  of  diseases,  such  as  heart  failure,  depression,asthma, inflammation and erectile dysfunction (Conti et al. 1995, Torphy 1998, Mehats et al.2002,  Rotella  2002,  Halena  and  Siegel  2007).  Several  PDE  inhibitors  are  currently  on  the

market, e.g. sildenfil (Viagra®), tadalafil (Cialis®) and vardenafil (Levitra®); all of these areinhibitors of PDE5.

2.2.6 Bone metabolism and PDE inhibitors

Several  studies  have  shown  that  selective  chemical  PDE  inhibitors  promote  osteoblasticdifferentiation and increase bone formation both in vivo and in vitro (Miyamoto et al. 1997,2003, Kinoshita et al. 2000, Wakabayashi et al. 2002, Scutt et al. 2004). Most of these effectsare likely to be mediated by elevated cAMP levels, although cAMP­independent effects havealso  been  reported  (Rawadi  et  al.  2001).  In  osteoblasts,  for  example,  cAMP  produced  inresponse  to  PTH  or  PGs  regulates  osteoblastic  differentiation  (Kumegawa  et  al.  1984,Farndale et al. 1988, Partridge et al. 1994, Ishizuya et al. 1997). There are corresponding datathat  show  that  administration  of  PTH  or  PGs  also  leads  to  increases  in  cancellous  bone

35

volume in animal models (Jee et al. 1985, 1987, High 1987, Finkelstein et al.1994, Whitfieldand  Morley 1995, Reeve 1996).  In rat bone marrow culture, XT­44 (a PDE4  inhibitor)  androlipram  stimulated  mineral­nodule  formation,  whereas  it  inhibited  osteoclast­like  cellformation  in mouse bone  marrow culture  (Waki et al.  1999). Therefore, PDE  inhibitors  areconsidered to be promising candidates for anti­osteoporosis drug therapies. A PDE4 inhibitor,XT­611,  has  been  shown  to  inhibit  osteoclast  multinucleation  or  maturation  by  actingsynergistically on osteoclast progenitors with prostaglandin E2 (PGE2) secreted from stromalcells,  but not by  influencing  the cell­to­cell  interaction  between stromal  cells and osteoclastprogenitors (Yamagami et al. 2003). Rolipram has augmented the inhibitory effects of PGE2on  osteoclast  progenitor  cell  viability;  thus  combined  treatment  with  PGE2  and  rolipramsuppresses osteoclast formation by directly reducing osteoclast progenitor cell viability (Parkand  Yim  2007). By  contrast,  PDE  inhibitors  have  been  shown  to  stimulate  osteoclastformation  by  inducing  receptor  activator of NF­{kappa} B  ligand  (TRANCE) expression  inmice  calvarial  osteoblasts  and  UAMS­32  cells  (Takami  et  al.  2005).  Pentoxifylline  androlipram  have  reportedly  been  to  increase  bone  mass  in  animals given  these drugs  in  dailyinjections (Kinoshita et al. 2000). PTH as well as PDE inhibitors induced osteoclast formationin  a  mouse  co­culture  system.  Injections  of  PTH  to  animals  once  per  day  produced  a  netanabolic effect in vivo (Tam et al. 1982, Fu et al. 2002). Continuous exposure to PTH leads toan increase in both bone formation and bone resorption, with a net loss of bone mass (Tam etal.  1982). PDE inhibitors  may  induce bone  loss  if  administered continuously  (Takami et al.2005).

2.3 Other signal transduction pathways associated with PDEs

PDE  isoenzymes  are  targets  of  intricate  regulatory  mechanisms  that  integrate  differentsignalling  pathways,  function  in  feedback  to  modulate  signalling  pathways  and  serve  aseffectors of signalling pathways. An additional property that only recently is being explored isthe subcellular distribution of PDEs and their  involvement in establishing microdomains andchannelling  of  signals.  Future  pathway  studies  are  anticipated  to  shed  light  on  the  role  ofPDEs in cyclic nucleotide signal transduction and connections between other signal pathways.

2.3.1 Exchange protein directly  activated by cAMP (EPAC) and PDEs

In  the  heart,  signalling  events  such  as  the  onset  of  cardiac  hypertrophy  are  influenced  bymuscle­specific mAKAP signalling complexes that target PKA, the EPAC and PDE4 (Dodge­Kafka  et  al.  2005).  Mediation  of  signalling  events  by  AKAPs  might  also  have  a  role  inlipolysis  in  adipocytes.  cAMP  either  stimulates  or  inhibits  PKB  phosphorylation,  actingthrough different effectors, in rat adipocytes. Hence, a cAMP pool that activates EPAC and isregulated by PDE3B and PDE4 appears to mediate a decrease in PKB phosphorylation. PKB

36

phosphorylation, and therefore activation, in adipocytes is regulated by an orchestrated cross­talk  between  insulin  and  cAMP  pathways,  where  PDE3B  seems  to  play  an  important  role(Zmuda­Trzebiatowska et al. 2007).

2.3.2 Mitogen­activated protein kinase (MAPK) and PDEs

PDEs  have  been  shown  to  be  involved  in  MAPK  signalling;  e.g.  pentoxifylline,  a  non­selective PDE  inhibitor,  induces osteoblastic differentiation  in pluripotent mesenchymal celllines  C3H10T1/2  and  C2C12.  Pentoxifylline  treatment  enhances  osteoblast  differentiation,which  occurs  via  signalling  by  MAPK  cascades  and  is  independent  of  PKA  activation(Rawadi  et  al.  2001).  Rat  mesangial  cells  (MCs)  possess  functionally  compartmentalizedintracellular pools of cAMP  that are differentially    regulated by  cAMP PDEs.  Inhibition ofPDE3,  but  not  PDE4,  has  been  demonstrated  to  suppress  MC  mitogenesis  (Tsuboi  et  al.1996). MC mitogenesis  is  regulated  through "negative cross­talk" between cAMP PKA andextracellular  signal­regulated  kinase  (ERK)  signalling.  PDE3  inhibitors  act  by  suppressingERK activation. Other targets of cAMP signalling directed by PDE3 include cyclins D, E andA, and the cell cycle inhibitor p21 (Cheng et al. 2004).

2.3.3 PKC and PDEs

PDEs have been reported to be  involved  in the non­canonical Wnt signalling pathway. Wntbinds to a Frizzled receptor and possibly also the co­receptor, Knypek or Ror2. This can leadto activation of a pathway involving calcium/CaM­dependent kinase II  and protein kinase C.Frizzled  may  activate  heterotrimeric  GTP­binding  proteins,  leading  to  activation  ofphospholipase C (PLC) and PDE. Frizzled can also recruit Dishevelled,  leading to activationof small GTP­binding proteins, such as Rho and  Cdc42 (Veeman et al. 2003). PKC, cAMP,PLC,  Ca2+  and  PDEs  have  been  shown  to  be  involved  oocyte  maturation  in  amphibians(Toranzo  et  al.  2007).  The  roles  of  PKC  and  PDE  have  been  demonstrated  in  ­opioidagonist­induced  reductions  in  cAMP  levels    in the  isolated  iris­ciliary  bodies  of  the rabbit(Dortch­Carnes  and  Potter  2002).  In  addition,  ­opioid  receptor  stimulation  in  rat  heart  hasshown PDE4 activation by PKC (Bian et al. 2000).

2.4 Bioactive peptides

During the  last decade,  fundamental studies have opened a new field of research concerningbioactive  or  biogenic  substances derived  from  foods  (Meisel  2001).  Proteins  and  peptidesfrom food have been found to be physiologically active or bioactive either in a direct mannerthrough their presence in the undisturbed food itself or after their release from the respectivehost proteins by hydrolysis in vivo or in vitro, e.g. in cheese ripening (Rizzello et al. 2005a),

37

food fermentation (Fitzgerald and Murray 2006, Korhonen and Pihlanto 2006) or enzymaticreactions  in  the gut after  the  ingestion of  foods containing precursor proteins  (Meisel et al.2003, Hernandez­Ledesma et al. 2004). Bioactive peptides have been found, among others, inmeat, fish, egg, soy and grain products, but thus far most have been isolated from milk­basedproducts. Functionally bioactive peptides are usually unknown sequences 2­5 aa in length andhave  a  proline  residue  in  the  C­terminus  (Curtis  et  al.  2002).  Two  human  di­/tri­peptidetransporters have been  identified; hPepT1  is expressed both  in the small  intestine and  in theproximal tubule, whereas hPepT2  is expressed only  in the proximal tubule (see Nielsen andBrodin  2003,  Steffansen  et  al.  2004).  The  cellular  mechanisms  and  signalling  of  bioactivepeptides remain obscure.

Different health effects have been attributed to food­derived peptides. Antimicrobial peptideshave been recognized in many protein hydrolysates, particularly in milk. The most studied arethe  lactoferricins,  which  are  derived  from  bovine  and  human  lactoferrin  (Kitts  and  Weiler2003).  In  addition,  antibacterial  peptides  have  also  been  recognized  in  s1­casein  and  s2­casein  (Rizzello  et  al.  2005b,  McCann  et  al.  2006).  Antimicrobial  peptides  operate  againstdifferent Gram­positive and Gram­negative bacteria, yeast and filamentous fungi. The ACEsare  involved  in  the  regulation  of  the  specific  maturation  or  degradation  of  a  number  ofmammalian bioactive peptides. Since the first detection of exogenous ACE inhibitors in snakevenom  (Ondetti  et  al.  1977),  a  large  number  of  ACE  peptides  have  been discovered  in  thedigestion of different food proteins, being established particularly in milk but also in fish andmeat. ACE inhibitory peptides are generally short­chain peptides, often carrying polar aminoacid  residues  like  proline  (Fuglsang  et  al.  2003)  The  hypertensive  and  immunomodulatorpeptides Val­Pro­Pro and  Ile­Pro­Pro,  for  example,  can be  released  from precursor proteins

­casein  and  ­casein)  by  enzymes  from Lactobacillus  helveticus  (Hata  et  al.  1996,Korhonen and Pihlanto 2003). The antihypertensity activities of bioactive peptides on bloodpressure have been shown in several studies (e.g. Mitsui et al. 2004, Miguel and Aleixandre2006). Hypocholesterolemic effects have been reported for casein and whey­derived peptides,in  addition  to  peptides  from  soy  protein  (Hori  et  al.  2001).  Antithrombotic  and  antioxidantactivities of bioactive peptides (Kitts and Weiler 2003) have been shown, as has enhancementof  mineral  absorption  and/or  bioavailability  in  rats  (Erba  et  al.  2002)  and  in  the  humanintestine (Chabance et al. 1998, Meisel et al. 2003). Immunomodulatory peptides can improveimmune  cell  functions,  measured  as  lymphocyte  proliferation,  natural  killer  cell  activity,antibody  synthesis  and  cytokine  regulation  (Horiguchi  et  al.  2005,  Fitzgerald  and  Murray2006).  Furthermore,  immunomodulatory  peptides  might  reduce  allergic  reactions  in  atopichumans and augment mucosal  immunity in the gastrointestinal tract (Korhonen and Pihlanto2003). Cytochemical studies have given increasing evidence that bioactive peptides modulatethe  proliferation,  differentiation  and  apoptosis  of  various  cell  types.  Some  milk­derivedpeptides,  for  instance,  have  been  shown  to  trigger  apoptosis,  mainly  in  malignant  cells,whereas normal cells seem to be less susceptible (Meisel and Fitzgerald 2003). Peptides havealso  been  observed  to  have  opioid  activities;  for  example,  opioid  receptor  ligands  withagonistic activity originate from different milk proteins and exert naloxone­inhibitable opioid

38

activities  in  both  receptor  studies  and  bioassays  (Kayser  and  Meisel  1996,  Mullally  et  al.1996).

2.4.1 Bioactive peptides in bone

Effects of bioactive peptides have been studied in dental/orthopaedic biomaterials as well asin basic bone biology. Proactive or new­generation biomaterials achieve specific, timely anddesirable  responses  from  surrounding  cells  and  tissues.  Osteoblast  adhesion  is  a  crucialprocess at bone implant  interfaces before multiple osteoblast functions (such as proliferationand mineralization), which occur after cell  adhesion and are necessary  for  the success of animplant  (Dee  1996a).  In  the  past  two  decades,  many  cell  adhesion  molecules  have  beendiscovered,  and  their  functions  in  cell  morphology,  locomotion,  mitosis,  cytokinesis,phagocytosis  and  maintenance  of  cell  polarity  have  been  investigated  (Pavalko  and  Otey1994,  Haass  et  al.  2005,  Turowski  et  al.  2005). A  key  finding  remains  that  the  amino  acidsequence  arginine­glycine­aspartic  acid  (RGD)  in  fibronectin  serves  as  a  primary  cellattachment  cue.  The  RGD  sequence  is  expressed  in  many  extracellular  matrix  (ECM)molecules, and the responsiveness of the RGD sequence for all attachments has been shownnumerous  times.  The  small  synthetic  peptides  (often  hundreds  of  daltons)  that  contain  theamino  acid  sequence  RGD  can  apparently  mediate  cell  attachment  similarly  to  theirconsiderably  larger parental  molecules (100 000 Daltons). According  to Craig et al.  (1995),the  potential  exists  to  develop  RGD­based  therapeutics  that  function  either  as  agonist,  topromote the interaction of cells and tissues with artificial matrices or as antagonists to controlthe  nature  of  cell­cell  and  cell­ECM  interactions.  The  coating  of  biomaterials  with  RGDpeptides offers an attractive strategy for controlling the cell­material interface and achieving abioactive  implant (Kantlehner et al. 1999, Verrier et al. 2002). Peptides have been shown tosupport and enhance osteoblast adhesion (Dee et al. 1998, Schuler et al. 2006), proliferation(Huag et al. 2003, Benoit and Anseth, 2005) and mineralization (Dee et al. 1996b). Based onthe RGD structure,  synthetic RGD­mimetic compounds have  been designed. These mimeticcompounds are effective inhibitors of bone resorption in vitro and in vivo (James et al. 1996,Engleman et al. 1997).

In addition,  the effect of other  small peptides  (without RGD sequence),  such as  isoleucine­proline­proline (IPP) and valine­proline­proline (VPP), have been studied  in bone cells. IPPand  VPP,  fermented  from  milk  by Lactobacillus  helveticus,  increase  activation  of  boneformation  in osteoblast culture, but do not have an  impact on osteoclast activation and boneresorption in vitro (Narva et al. 2004). L. helveticus­fermented milk has been demonstrated toprevent bone  loss  in ovariectomized  rats, but  this phenomenon could  not be  seen  followingsynthetic VPP treatment (Narva et al. 2007).

39

3. Aims of the study

This study focused on the areas of bone cell biology and signalling transduction. PDEs are animportant part of signal transduction in the cell, but very little  is known about them in bonecells.  The  inhibition  of  some  PDEs  could  have  therapeutic  potential  in  the  treatment  ofosteoporosis. Short bioactive peptides have been shown to have modulatory effects on bonebiology.

The aim of  this  study was  to  identify PDE  families  in  human osteoblasts and osteosarcomacells as well as to investigate the role of PDEs and bioactive peptides on the activation and thedifferentiation of human osteoblast cells.

Specific aims were as follows:

1) To identify the expression of PDE families in human osteoblasts.2) To investigate the effects of PDE inhibition on osteoblast differentiation and function

with RNA interference (RNAi) and pharmacological inhibitors.3) To  analyse  the  expression  and  regulation  of  different  PDEs  during  osteoblast

differentiation at the mRNA and enzyme levels.4) To examine the effects of bioactive peptides on gene expression in the osteoblast.

40

4. Materials and methods

4.1 Cell culture

4.1.1 Osteosarcoma cell lines (I, III, IV)

Human and rat osteosarcoma cell  lines, widely used as models  for osteoblasts, were used instudies I, III and IV. The human osteosarcoma MG­63 (III), which is considered to represent amature osteoblastic phenotype was the primarily used cell  line, whereas SaOS­2 cells (I, III)were mainly used to confirm results obtained with MG­63 cells (III) or NHOst (I). Both celllines were purchased from the American Type Culture Collection. MG­63 cells were culturedin α­MEM  supplemented  with  10%  foetal  calf  serum  and  SaOS­2  cells  in  McCoy`s  5A,supplemented  with  12.5%  foetal  calf  serum,  with  antibiotics  (50  IU/ml  penicillin  and  50µg/ml  streptomycin)  in a  humidified atmosphere  with 5% CO2 at 37°C. UMR­106  (IV)  ratosteosarcoma cells (American Type Culture Collection) were grown in Dulbecco’s modifiedEagle’s medium (DMEM), supplemented with 10% foetal calf serum (FCS), 50 IU penicillinand 50  g/ml streptomycin at 37°C in a humified atmosphere with 5% CO2 on 96–well plates.The cells were seeded at 1 x 104 cells/cm².

4.1.2 Human osteoblasts (I)

Normal human osteoblasts (NHOst) and all the cell culture reagents for the NHOst cells werepurchased from Cambrex, BioWhittaker. Two strains of NHOst cells isolated from bone chipsfrom  female  donors  were  used.  The  cells  were  seeded  at  a  density  of  5000  cells/cm2  andcultured in osteoblast growth medium supplemented with 10% FCS, 50 µg/ml ascorbic acid, 5mM  ß­glycerophosphate,  200  nM  hydrocortisone­21­hemisuccinate  and  0.1%gentamin/amphotericin­B solution, at 37ºC in a 5% CO2/95% air atmosphere on plastic Petridishes.

4.1.3 Human mesenchymal stem cells (hMSC)­derived osteoblasts (II, IV)

Human  bone  marrow­derived  mesenchymal  stem  cells  (hMSC,  Cambrex)  were  cultured  inmesenchymal  stem  cell  growth  medium  with  mesenchymal  cell  growth  supplements(Cambrex)  at  37°C  in  a  humified  atmosphere  with  5%  CO2  on  100­mm  plastic  dishes  at  aseeding  density  of    3100  cells/cm2. The  differentiation  into  osteoblasts  was  performedaccording  to  the  instructions  of  the  supplier  after  the  cultures  became  50%  confluent.  Thedifferentiation  medium  consisted  of  ­MEM  supplemented  with  10%  foetal  bovine  serum,10­8  M  dexamethasone  (Dex),  50  g/ml  L­ascorbic  acid,  10  mM  –glycerophosphate,  100U/ml penicillin and 100 U/ml streptomycin and was changed every 3­4 days.

41

4.2 PDE inhibitors

The PDE chemical inhibitors used in the studies are listed in Table 3.

TABLE 3. PDE inhibitors used in Studies I­III.

Family Inhibitor Study

PDE1 8­MMX (8­methoxymethyl­3­isobutyl­1­methylxanthine) I, III

PDE2 EHNA I, III

PDE3 Cilostamide (N­cyclohexyl­N­methyl­4­[1,2­dihydro­2­oxo­6­quinolyloxy] butyramide)

III

PDE3 Milrinone (6­methyl­2­oxo­5­pyridin­4­yl­1H­pyridine­3­carbonitrile)

I

PDE4 Rolipram (4­[3­{cyclopentyloxy}­4­methoxyphenyl]­2­pyrrolidinone)

I, II, III

PDE7 BRL 50481 (3­(N,N­dimethylsulfonamido)­4­methyl­nitrobenzene)

II

PDE3 cGMP I, III

non­specific PDEinhibition (PDE1­7,10­11)

IBMX I, III

4.3 Primers (I, II, III, IV)

The primers were used as presented in Table 4.

TABLE 4. Primers used in Studies I­IV. Forward primer (F), Reverse Primer (R).cDNA  Sequence StudyPDE1A F: AAAATGGGGATGACAAAAAAGAA

R: GATATTCATTTCTTCTTCTTGCATI, III

PDE1B F: ATGCAGGATGATGAGATGAAR: CGGAAGAATTCCTCCAT

I, III

PDE1C F: AGACAGGGTGACAGAGAAGCR: TGGCCTTCTCCTCTTTGGGT

I, III

PDE3A F: TCACCTCTCCAAGGGACTCCTR: CAGCATGTAAAACATCAGTGGC

I, III

42

PDE4A F: CTCGCACAAGTTCAAAAGGATGR: GCCTCCAGCGTAATCCGACA

I, III

PDE4B F: GACATTGCAACAGAAGACAAG TCCCCR: ACTCAAGTAACTGAAGGCCAGGTGG

II

PDE4B F: AGCTCATGACCCAGATAAGTGR: ATAACCATCTTCCTGAGTGTC

I, III

PDE4C F: ACCTCAGCGCCAAGCAGCR: TGGGTGGGAAAGTGAAGCAGG

I

PDE4C F: ACACTGAACTCCTGTCCCCTGAAGR: GATGTGACTCAAGAGTGACCACTGG

III

PDE4D F: CCCTGGACTGTTATCATGCAR: TGATTGGACACACCAGGATG

I, III

PDE7A F: GGACGTGGGAATTAAGCAAGCR: TCCTCACTGCTCGACTGTTCT

I, II, III

PDE7B F: TGCACAGGACAGGCACTTTAR: CTTCTGTGCTGCCTGGGCAA

I, II, III

PDE8A F: GCC TGT TTC CTG GACAAA CR: GCA TTA CGG ACA ACT CTT CTC

II

PDE8A F: CAGCCAGAGACGACACTCTTCCATR: ATGATCTTAGCGTTGACTCGGAGC

III

PDE10A F: GTGTGGTGCAGATGGTCAACR: GGGTGAATTGCATGAGACCT

III

PDE11A F: ATCTGCCTCAGTATCCCCCTR: CCACCAGGAAAAGAGAGCAG

III

GAPDH F: AAGGCTGGGGCTCATTTGR: CCACCAGGAAAAGAGAGCAG

III

Beta­actin F: CATCGTCACCAACTGGGACGACR: CGTGGCCATCTCTTGCTCGAAG

I

Beta­actin F: AGGCCAACCGCGAGAAGATGACCR: GAAGTCCAGGGCGACGTAGCAC

II, IV

Osteocalcin  F: ATGAGAGCCCTCACACTCCTCGR: GTCAGCCAACTGGTCACAGTCC

II, IV

Caspase­8 F: AGGAGGAGATGGAAAGGGAACTTR: ACCTCAATTCTGATCTGCTCACTTCT

II, IV

CREB­5 F: GCTTTGGTGCTTTTCTCCAGR: GGTGACACCACAGCACAAAC

II, IV

Beta­catenin  F: TTCTGGTGCCACTACCACAGCR: TGCATGCCCTCATCTAATGTC

II

Vitamin Dreceptor

F: CCAGTTCGTGTGAATGATGGR: CCTTTTGGATGCTGTAACTG

II

BMP­1 F: AGGCCCACTTCTTCTCAGAAAAGR: GAAGAGAGCACCAGTTCTGTCCG

II

BMP­2 F: GGAGAAGGAGGAGGCAAGR: GACACGTCCATTGAAAGAGC

IV

BMP­5 F: AAGAAGACAAGAAGGACTAAAAATATR: GTAGAGATCCAGCATAAAGAGAGGT

II, IV

PTHrp F. GTCTCAGCCGCCGCCTCAAR: GGAAGAATCGTCGCCGTAAA

IV

T7­promotor+ PDE7

F: TAATACGACTCACTATAGGTTCAAGAAGATTACR: TAATACGACTCACTATAGGTTTCACTCCACT

II

43

T7­promotor + PDE8A

F: TAATACGACTCACTATAGGAAGAAGGTAGCAGR: TAATACGACTCACTATAGGTTCATTGTACGAG

II

4.4 cAMP analyses (I, II, III)

Measurement  of  cAMP  accumulation  was  determined  with  a  luminometric  cAMP  assay(cAMP­Glo TM Assay, Promega) two days after transfections with PDE7 siRNAs (II) or usingthe  method  of  Frandsen  and  Krishna  (1976),  as  described  earlier  (Ahlström  et  al.  2000).Briefly,  the  cells  were  grown  in  24­well  plates  without  treatment  (I)  or  treated  with  eithervehicle (ethanol) or Dex (1­100 nM) for 48 h (III). The cells were then induced with forskolinfor  7  min  at  37°C  and  rinsed  twice  with  ice­cold  phosphate­buffered  saline  (PBS)  (III)  orwashed  with  Hanks  Balanced  Salt  Solution  (HBSS)  (I).  For  determination  of  basal  cAMPlevels,  cells  were  incubated  at  37°C  for  30  min  before  washing  with  ice­cold  PBS  (III)  orHBSS (I). cAMP was extracted from the cells with 96% ethanol at ­18°C for 3 h. The extractwas evaporated with N2 gas and dissolved in assay buffer (0.05 M sodium acetate, pH 6.2). Ifnecessary, the sample was further diluted with assay buffer, and the cAMP concentration wasdetermined by radioimmunoassay with iodinated cAMP as tracer.

4.5 Q­sepharose chromatography (I, III)

Five confluent cultures of either NHOst, SaOS­2 or MG­63 cells, grown on 100­mm plasticdishes were washed twice with PBS buffer, harvested with a cell scraper into 6 ml of ice­coldhomogenization  buffer  (buffer  A),  containing  20  mM  Bis­tris,  pH  6.5,  5  mMmercaptoethanol, 0.1% of a protease inhibitor cocktail (PIC), consisting of 0.5 mg/ml each ofbestain, pepstatin A, aprotinin,  leupeptin, E­64 and 12.5 mg/ml AEBSF. The harvested cellswere  then  homogenized  on  ice  by  ten  passages  with  a  Teflon/glass  homogenizer.  Thehomogenization was repeated three times. The homogenate was then centrifuged for 20 minat 20 000 g. Five millilit of the supernatant was diluted with 20 ml of buffer B, containing 20mM  Bis­tris,  pH  6.5,  0.1  M  sodium  acetate,  0.02%  PIC  (v/v),  0.1  mM  EDTA,  1  mMbenzamidine  and  1  mM  mercaptoethanol,  filtered  through  a  0.22­µm  syringe  filter  andapplied  to  a  column  (5­ml  bed  volume)  of  Q­sepharose  High  Performance  (AmershamPharmacia Biotech) previously equilibrated with buffer B. After washing the column with 10bed volumes, the PDE activities were eluted with a 0.1­1.0 M linear sodium acetate gradientin  buffer  B,  at  a  flow  rate  of  2.5  ml/min.  Fractions  of  2  ml  were  collected  into  tubescontaining 50 µl of  5% bovine serum albumin (BSA) and assayed for either cAMP­PDE orcGMP­PDE activity, as described below.

44

4.6 Enzyme activity analyses

4.6.1 Assay of PDE activity (I, II, III)

PDE  activity  was  assayed  according  to  the  method  described  by  Thompson  and  Appleman(1971). Samples  from either Q­sepharose elutions or whole­cell homogenates were added tothe incubation buffer (40 mM Tris­HCl, pH 8.0, 0.1% PIC, 0.05% BSA, 5 mM MgCl2, 1 mMEGTA, 0.5 µM cAMP and 3H –cAMP (100000 cpm/tube) to give a final reaction volume of

200 µl. For the cGMP­PDE assay, cAMP was replaced with 0.25 µM cGMP and 3H –cGMP.The  samples  were  incubated  at  34°C  for  30­60  min,  followed  by  boiling  for  1  min.  Aftercooling on ice, 50­100 µl of Crotalus atrox snake venom nucleotidase (1 mg/ml, Sigma) wasadded  and  incubated  at  34°C  for  10  min.  Then  500 µl  of  a  1:3  slurry  of  AGI­X8  resin(BioRad) was added, and the tubes were mixed and centrifuged  for 5 min at 8000 rpm. Theradioactivity of the supernatant was determined with a β­counter. Selective inhibitors used fordifferent PDE types are summarized in Table 3.

4.6.2 Assay of ALP activity (II)

Bone­specific alkaline phosphatase (bALP) activity was measured with a immunoenzymetricassay  (OCTEIATM  Ostase®  BAP,  Immunodiagnostic  Systems).  Prior  to  measurements,  thecells  were washed with PBS and after addition of 300  l of Tris –Triton  (0.1 M Tris–base,0.2% Triton X­1000), the cells were frozen at ­70 C. After thawing, the cells were centrifuged(12000 g, 5 min). The supernatant was used for analysis. The ALP activity was corrected withtotal protein content.

4.7 PDE silencing (II)

PDE  silencing  was  performed  with  siRNAs,  and  the  PDE7  selective  inhibitor  BRL  50481(Table 3) was  mainly used  to confirm results obtain with PDE7  siRNAs. The short double­stranded RNA (dsRNA) mixtures for PDE7 and PDE8 inhibition were generated from PDE7and PDE8 PCR products. Primers with appended T7 promoters sequenced for the PDE7 PCRproducts  were  designed  to  amplify  both  PDE7A  and  PDE7B.    PDE8  PCR  products  weredesigned to amplify only PDE8A. The BLAST database program (NBCI) was used to confirmthe specificity of  the  target  sequences. The  large  dsRNAs were  first  synthesized by  in  vitrotranscription  from the DNA templates. The  large (150­1500 bp) dsRNAs were processed byShortCut RNase III digestion in the presence of manganese, which produced a heterogeneouspopulation of short (18­25 bp) siRNAs. For generation of siRNAs, we used a ShortCut RNAikit (New England BioLabs). After 12­14 days of differentiation, the osteoblast phenotype was

45

confirmed  by  staining cultures and  control cultures with a  histological alkaline phosphatasekit  (Sigma).  At  this  stage,  the  medium  was  changed  to  a  low  serum  (2.5%),  antibiotic­freemedium, and 24 h later, the siRNA mixtures were transfected into the cells by Lipofectamin2000  (Invitrogen).  Control  cells  were  simultaneously  treated  with  lipofectamin  withoutsiRNAs.  The  transfection  time  varied  between  6  and  24  h,  and  the  cell  culture  was  thenchanged  to  a  new  differentiation  medium.  The  transfection  efficiency  was  determined  withBLOCK­iT™  Fluorescent Oligos (Invitrogen) according to the manufacturer`s instructions.

4.8 Isoleucine­proline­proline (IPP), valine­proline­proline (VPP) and leucine­lysine­proline(LKP) tripeptides (IV)

The  peptides  were  synthesized  by  9­fluorenylmethyloxycarbonyl  (Fmoc)  chemistry  andpurified using HPLC reverse phase columns. Peptides were synthesized  in  the Core Facilityfor  Synthetic  Peptides,  Division  of  Biochemistry,  Department  of  Biological  andEnvironmental   Sciences, University of Helsinki, Finland. Peptides were dissolved  in  sterilewater prior to administration.

4.9 Cell proliferation analyses (II, IV)

The proliferation rate was assayed with a luminometric ATP detection assay (ATPlite, PerkinElmer)  48  h  after  the  siRNA  transfections  (II),  with  multilabel  counter  (Victor  3,  PerkinElmer) or measured the activity for DNA synthesis (IV) with a Cell Proliferation ELISA kit(Roche  Diagnostics).  The  5­bromo­2’­deoxyuridine  (BrdU)  binds  to  the  newly  synthesizedcellular DNA. BrdU incorporation was measured with a spectroscopic plate reader at 450 nm(Multiskan Ex, Thermo Labsystems).

4.10 Mineralization analyses (II)

Mineralization  was  assessed  by  Alizarin  Red  S  staining  in  cells  grown  on    96­well  platesusing the method described by Stanford et al. (1995). After staining, the dye was eluated with10% cetylpyridinium chloride in 10 mM sodium phosphate (pH 7.0, 15 min, RT). Absorbancewas measured with a spectroscopic plate reader at 590 nm.

46

4.11 mRNA expression analyses

4.11.1 RNA isolation (I­IV)

RNA of the osteoblastic cells was extracted using the QuickPrep Micro mRNA purificationkit  (Amersham  Pharmacia  Biotech)  or  the  RNeasy  Protect  mini  kit  (Qiagen  GmbH).  TheRNA concentration was measured at 260 nm with a  spectrophotometer. The  integrity of  theRNA was confirmed by 1.2% formaldehyde agarose gel electrophoresis.

4.11.2 cDNA microarrays (II, IV)

The cDNA microarray hybridizations were performed at the Finnish DNA Microarray Centreof  the Turku Centre for Biotechnology. The expression profiles of osteoblasts differentiatedfrom  hMSC  were  compared  after  a  48­h  treatment  with  35  nM  PDE7  or  PDE8A  siRNAmixtures (III), or after a 24 h treatment with 50  M IPP, VPP or LKP (IV), with Hum16­Kprotocol.  The array  contains 16  000  human  gene  probes  (accession  number:  A­MEXP­557,http://www.ebi.ac.uk/ arrayexpress/). Samples were labelled with FluoroLink Cy­3­dUTP andCy­5­dUTP  (Amersham  Biosciences)  using  20  g  of  total  RNA  for  direct  labelling  andhybridized using an augmented reference design (Kerr and Churchill 2001). Discrete  imagesfor  Cy­3  and  Cy­5  dyes  were  obtained  using  a  Scan  Express  laser­scanning  microscope(Packard  BioSciences)  and  gene  transcript­levels  were  determined  from  the  fluorescenceintensities  of  scanned  data  image  files  with  the  QuantArray  Microarray  Analysis  software(Packard  BioSciences).    T­test  and  gene­expression  profiles  were  analysed  by  KensingtonDiscovery  Edition  2.0  software.  At  least  a  1.6  fold  difference  in  expression  was  used  incontrol/treatment  comparison.  The  cDNA  microarray  experiment  is  available  at  accessionnumber E­MEXP­1021 (II) and E­MEXP­885 (IV) (http://www.ebi.ac.uk/arrayexpress/). Thegene  expression  data  for  IPP­  treated  osteoblasts  were  analysed  by  principal  componentanalysis  (PCA)  to  determine  the  main  components  affected.  Results  were  consideredsignificant at a 95% significance level (p < 0.05) (IV).

4.11.3 cDNA synthesis (I, II, III, IV)

cDNA  was prepared  from 1­2 µg of  total RNA or  mRNA  by  reverse  transcription with  thefollowing buffer and conditions: RT buffer (Promega) containing 50 mM Tris­HCl (pH 8.3)75 mM KCl, 3 mM MgCl2 and 10 mM DTT, 13 U RNAguard, Rnase inhibitor (AmershamPharmacia Biotech), 0.9 mM each of dATP, dCTP, dGTP and dTTP, 15 pmol  Oligo(dT) 15primer    (Promega)  and  100  U  M­MLV­reverse  transcriptase  primer  (Promega)  in  a  finalvolume of 25 µl, with incubation for 1 h at 37°C.

47

4.11.4 Semi­quantitative reverse transcriptase polymerase chain reaction (RT­PCR) (I, III)

PCR amplifications were conducted  in 30­µl aliquots containing 2­4 µl of cDNA, 1x buffer(Finnzymes), comprising 10 mM Tris­HCl, pH 8.8 at 25°C, 1.5 mM MgCl2, 50 mM KCl and

0.1% Triton X­100, 0.2 mM each of dATP, dCTP, dGTP and dTTP, 0.35 µM of each primerand 1 U of DyNAzyme DNA polymerase (Finnzymes). Primers are  listed  in Table 4. Theamplification protocol was run on a PTC­100 thermocycler (MJ Research). After denaturationat 94°C for 4 min, the reactions underwent cycles of denaturing for 1 min at 95°C, annealingfor 90 s at 58­63°C and extension  for 2 min at 72°C. This was  followed by a  single step at72°C for 10 min. To ensure that the PCR was in the exponential phase, the conditions of thesemi­quantitative RT­PCR were optimised  for the number of PCR cycles and the amount oftotal RNA applied to the reactions  for the PDE subtypes/isoforms and for glyceraldehyde­3­phosphate  dehydrogenase  (GAPDH)  or  ­actin,  which  was  used  as  an  external  standard.Human brain RNA and a cocktail of RNA from rat spleen, brain, kidney and heart was usedto  prepare  cDNAs  used  as  positive  controls.  Negative  controls  included  samples  in  whichcDNA synthesis was performed in the absence of reverse transcriptase. Reactions in which nocDNA was added were also performed to check for possible non­specific products. Volumesof  10  µl  of  the  PCR­amplified  products  were  analysed  on  1.5%  agarose  gels  containingethidium  bromide.  Gel  results  were  quantified  by  AlphaDigiDoc™   System  1000  (AlphaInnotech Corporation). The results were analysed as a ratio of integrated optical density of thePDE subtypes/isoforms to the external standard.

4.11.5 Quantitative Real­time PCR (qRT­PCR) (II, IV)

PCR  amplification  was  performed  in  a  real­time  quantitative  PCR  engine  Mx3000P(Stratagene). The PCR reactions contained cDNA templates and Brilliant SYBR Green QPCRMaster  mix  kit  (Stratagene)  in  a  25­µl  reaction  volume  containing  200  nM  of    each  gene­specific  primer  (Table  4).  Fluorescence  data  were  collected  during  the  annealing  step  andanalysed with Mx3000P software. Amplification was obtained by denaturing at 95°C for 10min,  followed by 40 cycles of denaturing at 95°C for 30 s, annealing at 58°C for 1 min, andextension at 72°C for 30 s. The amplification cycles were followed by for 1 min at 95°C andfor  30  s  at  55°C,  and  data  for  dissociation  plots  were  collected  as  the  temperature  wasreturned  to  95°C.  cDNA  from  human  brain  total  RNA  (BD  Biosciences)  was  used  as  acalibration  standard,  and  the  data  were  normalized  with  amplification  data  of  ß­actin.  Allreactions  were  run  in  triplicate  with  three  dilution  concentrations,  and  the  mean  value wasused to calculate the ratio of target gene/ß­actin expression in each sample. Using the ratio inthe untreated sample as a standard (100), the relative ratio of treated sample was presented asthe relative expression level of the target gene.

48

5. Results and discussion

5.1 Identification of PDE families in human osteoblastic cells (I, III)

PDEs has been extensively studied in kidney, brain, heart, muscle and immune system cells ofhumans,  rats  and  mice  (Lugnier  2006),  but  very  little  is  known  about  them  in  bone  cells.There  are  no  reports  of    PDEs  in  human  osteoblasts.  In  Studies  I  and  III,  we  thereforeexamined the profile of PDEs in different types of osteoblastic cells, comprising SaOS­2 andMG­63 osteosarcoma cell lines and normal human osteoblast (NHOst) cells (Table 5). MG­63and  SaOS­2  are  osteoblast­like  cells  widely  used  as  osteoblast  models  in  in  vitro  studies(Murray et al. 1987, Rodan et al. 1987, Lajeunesse et al. 1990).

The cAMP PDE activity of NHOst, SaOS­2 and MG­63 cells separated by Q­sepharose ion­exchange chromatography was reflected in three peaks, A1­A3 in Figure 6 and, 1­3 in Figure7. The  first peak of NHOst, SaOS­2  (I)  and MG­63  (III)  cells consisted of  CaM­stimulatedPDE1 activity, which is sensitive to 8­MMX. This peak was also sensitive to cGMP, which isexplained  by  the  dual  substrate  nature  of  PDE1.  IBMX,  but  not  EHNA,  cilostamide,milrinone or rolipram, also affected this peak. The second peak activity of NHOst and SaOS­2cells  was  insensitive  to  the  addition  of  high  levels  of  several  tested  inhibitors.  IBMX,however,  inhibited  the  activity  of  the  second  peak.  PDE7,  the  cAMP­specific,  rolipram­insensitive PDE family, has been shown to be insensitive to most known PDE inhibitors, butrelatively sensitive to IBMX (Hetman et al. 2000), suggesting that the most  likely candidatefor peak A2 of both NHOst and SaOS­2 cell is PDE7.

TABLE  5. Summary  of  PDE  families  and  subtypes  expressed  in  human  osteoblasts  andosteoblast­like cells (+ identified, ­ not identified, / not determined).

PDE1 PDE3              PDE4  PDE7 PDE8  PDE10   PDE11

_______     _______        __________________     _______    ____      _____         ___1A 1C  3A  3B    4A   4B  4C  4D   7A  7B    8A   10A   11A

NHost +  +  +  ­    +   +  ­  ­    +  +      /     /       /

SaOS­2 ­  +  +  ­    +   +  +  +    +  +      /     /      /

MG­63 +  +  +  ­    +   +  +  +    +  +     +     +     +

49

Figure 6.   Elution profile of cAMP­PDE activity of NHOst and SaOS­2 cells: Cell extractswere  applied  to  a  Q­sepharose  column  and  eluted  with  a  linear  sodium  acetate  gradient(straight  line).  The  eluted  fractions  of  NHOst  (A)  and  SaOS­2  (B)  cells  were  assayed  forcAMP­PDE  activity  in  the  absence  (o)  and  presence  (g)  of  6 µM  cGMP,  and  in  thepresence of 20 U/tube CaM/10 µM CaCl2 (n). The eluted cGMP­PDE activities of NHOst(C) and SaOS­2 cells  (D) were assayed  in the absence (o) and presence (n) of 20 U/tubeCaM/10 µM CaCl2.

The  identity  of  the  third  peak  was  determined  by  assaying  sensitivity  to  cGMP,  and  PDEinhibitors indicated a different identity of A3 peaks of NHOst and SaOS­2 cells (Table 6). InNHOst  cells,  A3  was  shown  to  be  sensitive  to  milrinone,  a  PDE3  selective  inhibitor,  butinsensitive to rolipram, a PDE4 selective inhibitor. Peak A3 of SaOS­2 cells was sensitive torolipram, but relatively insensitive to milrinone. The high sensitivity of the third NHOst peakto cGMP further suggested that the NHOst peak A3 activity is produced by a member of thePDE3 family. The only detected cGMP PDE activity in both NHOst and SaOS­2 consisted oftwo closely eluting CaM­stimulated peaks. The first of these peaks mostly hydrolysed cGMP,but  the  second  peak,  corresponding  to  the  fractions  of  cAMP­hydrolysing  peak  A1,hydrolysed  both  cAMP  and  cGMP.  The  two  activities  of  MG­63  cells  (peaks  2  and  3)contained rolipram­sensitive PDE4 activity (Figure 7). EHNA, cilostamide and cGMP did notreduce the activity of these two peaks.

50

Figure  7.  Q­sepharose  elution  profile  of  cAMP­PDE  activity  present  in  MG­63  cells.  Cellextracts  were  applied  to  a  Q­sepharose  column  and  eluted  with  a  linear  sodium  acetategradient  (0.05–1.2 M). Fractions were collected  and  assayed  for  cAMP­PDE activity  in  thepresence (ο) or absence (•) of PDE inhibitors and cGMP. The concentrations of the inhibitorswere  as  follows:  8­MMX,  25  M;  EHNA,  10  M;  cilostamide,  5  M;  rolipram,  10  M;IBMX, 50  M. The concentration of cGMP was 5  M.

51

TABLE 6. Sensitivity of the eluted peaks towards selective PDE  inhibitors. The results aregiven as µM inhibiting half of the maximal PDE activity (IC50).

peak A1 peak A2 peak A3

inhibitor NHOst SaOS­2 NHOst SaOS­2 NHOst SaOS­2

8­MMX 11 8 >100 >100 >100 >100Milrinone >100 >100 >100 >100 0.13 10Rolipram >100 >100 >100 >100 >100 0.19IBMX ­ ­ 7 6 ­ ­cGMP 25 17 100 100 0.01 100

Peaks  2  and  3  were  also  sensitive  to  IBMX.  8­MMX  had  a  mild  inhibitory  effect  on  theactivity of peaks 2 and 3, possibly due to non­specific inhibition. IBMX inhibited the activityof  all  fractions.  The  relative  PDE1  and  PDE4  activities  of  MG­63  and  SaOS­2  cellhomogenates were also assayed (III). MG­63 cells contained higher  levels of PDE4 activitythan SaOS­2 cells.

According  to  biochemical  identification,  PDE1  is  one  of  the  major  PDEs  in  NHOst  andSaOS­2 cells  (I) and PDE4  is one of  the major PDEs  in MG­63 cells  (III). PDE expressionmay depend on the biological state and environment of the cell. For example, the transcript ofPDE4B2 is more abundant in the myometria of pregnant women than in non­pregnant women(Leroy et al. 1999). PDE3 activity was detected in NHOst and MG­63 cells, but SaOS­2 cellslacked  this  activity.  MG­63  and  SaOS­2  cells  are  malignant  cell  lines  that  are  in  differentstages  of  osteoblastic  differentiation,  which  could  in  part  explain  their  dissimilar  PDEprofiles.

5.2 Identification of PDE subtypes in human osteoblastic cells (I, III)

The presence of mRNA transcripts of all known PDE1, PDE3, PDE4 and PDE7 subtypes wasexamined  by  RT­PCR (Table 5). PDE1A  and PDE1C  transcripts were expressed  in  NHOstand MG­63 cells, whereas SaOS­2 cells only expressed the PDE1C subtype. Of the two PDE3subtypes, PDE3A and PDE3B, mRNA of the former was detected in all three cell types. Theexistence of mRNA transcripts of all known PDE4 subtypes (PDE4A­4D) (Bolger et al. 1993)was examined; PDE4A and PDE4B subtypes were found  in NHOst cells, whereas all PDE4subtypes  were  found  in  MG­63  and  SaOS­2  cells. HOSM­1,  an  osteosarcoma  cell  lineestablished from a human mandible, has demonstrated the expression of PDE4A, PDE4B andPDE4C, and rolipram has been shown to inhibit the proliferation of these cells (Narita et al.2003). Lower PDE4 activity  in NHOst than  in SaOS­2 and MG­63 cells may occur because

52

cells  expressed  only  PDE4A  and  PDE4B  subtypes.  The  expression  of  PDE4C  and  PDE4Dcould be involved in the proliferation and the malignant phenotype of osteosarcoma cells.

The  rolipram­insensitive/IBMX­sensitive  cAMP­specific  PDE  activity  detected  in  NHOst,SaOS­2  and  MG­63  cells  indicated  the  presence  of  PDE7.  PDE7  has  two  subtypes  inmammals; both mRNA transcripts (PDE7A and PDE7B) were detected in all studied cells. Inaddition, the presence of mRNA transcripts of the remaining cAMP PDEs (PDE8A, PDE10Aand PDE11A) was identified in MG­63 cells.

Profiles of  PDE  subtypes  in  different  osteoblastic  cells  have previously  been  studied  in  ratUMR­106  osteoblast­like  cells    (Ahlström  and  Lamberg­Allardt  1999,  2000),  mouseosteoblastic cell line MC3T3­E1 and  bone marrow stromal cell line ST­2 (Wakabayashi et al.2002).    In  mouse  cell  lines,  PDE1,  PDE2,  PDE3,  PDE4,  PDE7,  PDE8  and  PDE9  mRNAtranscripts have been identified. In rat UMR­106 osteoblast­like cells, the main PDE fractionsare PDE1 and PDE4, but PDE2, PDE3, PDE7 and PDE8 expression has also been detected.These findings are line with our results from human osteoblasts and osteosarcoma cells MG­63  and  SaOS­2.    PDE1,  PDE2,  PDE3,  PDE4,  PDE7  and  PDE8  may  be  the  major  PDEs  inosteoblasts,  but  profiles  and  subtypes  could  vary  during  the  differentiation  and  withenvironmental changes.

5.3 PDE7 and PDE8A silencing by siRNAs (II)

In  in  vivo  and  in  vitro  studies,  the  inhibition  of  PDE4  has  been  demonstrated  to  increaseosteoblastic differentiation and bone formation in rats and mice (Miyamoto et al. 1997, 2003,Kinoshita  et  al.  2000,  Wakabayashi  et  al.  2002,  Scutt  et  al.  2004),  but  the  effect  of  PDE7inhibition  on  bone  biology  has  not  previously  been  investigated.  At  the  beginning  of  ourstudy,  no chemical  inhibitors for PDE7 or PDE8 were available. siRNAs have provided thefield  of  molecular  biology  with  a  new  and  efficient  tool  to  efficiently  knock­down  specificgenes (Mousses et al. 2003, Bantounas et al. 2004, Semizarov et al. 2004). We  investigatedthe effect of specific  silencing of   phosphodiesterase subtypes PDE7A, PDE7B and PDE8Aon gene expression in hMSC­derived osteoblasts. For the silencing of specific PDEs, siRNAsgenerated  from  PDE7  and  PDE8  were  utilized.  Quantitative  real­time  PCR  was  used  toconfirm  the  specific  PDE  silencing.  PDE7  and  PDE8  inhibition  by  RNAi  significantlydecreased the gene expression of PDE7A, PDE7B and PDE8A. At the level of mRNA, PDE7and PDE8 inhibition by RNAi decreased the gene expression of PDE7A by 60­70%, PDE7Bby 40­50% and PDE8A by 30%.

Since the discovery of the RNAi method various genes involved in the cell cycle, apoptosis,differentiation  and  signalling  have  successfully  been  inhibited  in  different  mammalian  celltypes, in vitro and in vivo, to elucidate gene function, identify drug targets and treat a varietyof  diseases. β­catenin  expression  has  been  silencing,  leading  to  the  inhibition  of  growth  of

53

colon cancer cells (Verma et al. 2003). The reduction of   cyclin B1  levels  in HeLa cells bysiRNAs  caused  inhibition  of  proliferation  by  arresting  cells  in  the  G2  phase  and  inducingapoptosis  (Yuan  et  al.  2006).  RNAi  targeting  TGF­β1  significantly  ameliorated  theprogression of matrix expansion in experimental glomerulonephritis (Takabatake et al. 2005).In  addition,  the  gene  expression  of    PDE4B  and  PDE4D  has  been  reduced  with  syntheticsiRNAs  in  human  embryonic  kidney  (HEK293B2)  cells  (Lynch  et  al.  2005).  Silencingefficiency  of  siRNAs  is  dependent  on  the  sequence  of  siRNAs,  the  number  of  siRNAmolecules  and  the  cell  doubling  rate.  In  slowly  proliferating  or  non­dividing  cells,  siRNAsseem to have a longer gene silencing effect than in rapidly dividing cells (Pai et al. 2006). Forexample, siRNA activity lasts for 3­7 days in proliferating cells, yet persists for three weeksor more in terminally differentiated cells such as neurons (Omi et al. 2004). We used hMSC­derived osteoblasts in our study, when cells differentiated the rate of proliferation has becomelimited  (Amedee  et  al.  1994).  Our  results  show  that  specific  gene  silencing  with  the RNAimethod is useful for inhibiting the gene expression of PDE7 and PDE8 subtypes.

5.4 PDE7 silencing has potential for enhancing osteogenic gene expression anddifferentiation in hMSC­cells derived osteoblasts (II)

Gene  expression  profiling  was  performed  using  a  Hum­16K  cDNA  microarray.  The  arraycontains 16 000 human genes probes and  is designed  for genome­wide screening. The arraytherefore  contains  a  limited  number  of  known  key  factors  for  osteoblastogenesis  and  boneformation; Runx 2, Rankl and Lrp5 genes are missing. Treatment for 48 h with PDE7 siRNAsup­regulated 435 and down­regulated 60 loci (  1.6­fold difference). The respective numbersfor siRNA treatment of PDE8A were 38 and 96. Selected microarray results are displayed inTable 7. Quantitative real­time PCR results confirmed up­/down­regulation of a set of genesfrom the microarray results (Table 8). The qRT­PCR results were in line with the microarrayanalysis.  PDE7  silencing  up­regulated  the  gene  encoding  OCN,  a  non­collagenous  bone­specific  matrix  protein  secreted  by  normal  maturing  osteoblasts  (Price  et  al.  1976).  OCNexpression is highly up­regulated by 1,25(OH)2­D3 in MG­63 osteosarcoma cells (Mahonen etal.  1990,  Lajeunesse  et  al.  1991).  In  addition,  the  activation  of  the  cAMP/PKA  signallingpathway increases OCN mRNA and the secretion of OCN in rat osteoblast­like cells (Noda etal.  1988,  Theofan  et  al.  1989).  In  human  osteosarcoma  cells,  the  results  of  cAMP/PKAactivation on OCN expression are conflicting; OCN expression has been reported to be eitherup­ or down­regulated by compounds that elevate cAMP (Lajeunesse et al. 1991, Drissi et al.1997, Boguslawski et al. 2000).

54

TABLE 7. Effects of PDE7 and PDE8A RNAi silencing on gene expression in hMSC­derived humanosteoblasts. A set of interesting genes were selected from cDNA microarray results._________________________________________________________________________________________

Accession number Gene/protein    Fold­change___________________________________________________________________________Inhibition of PDE7

Up­regulatedEnzymes and metabolismH87471 Kynureninase (L­kynurenine hydrolase)            2.1DifferentiationAA779457 Bone morphogenetic protein 5           3.5

Catenin beta­like 1           2.0Calcium­sensing receptor           2.0Glutamate receptor metabotropic 3           3.1Tumour necrosis factor (ligand)superfamily member 10           3.2

ApoptosisN94588 CASP8 and FADD­like apoptosis regulator            2.1

Down­regulatedHormone and growth factor receptors

Vitamin D (1,25 dihydroxyvitamin D3 receptor)    ­1.7Fibroblast growth factor receptor          ­5.4

ApoptosisT95823 cyclin L2         ­1.8

Inhibition of PDE8A

Up­regulatedN55459 Metallothionein 1F (functional)             1.6R01682 Metalloprotease­related protein 1       1.6Down­regulatedApoptosisN94588 CASP8 and FADD­like apoptosis regulator             ­1.9DifferentiationAA045327 Osteoglycin (osteoinductive factor, mimecan)          ­1.5R56774 Bone morphogenetic protein 1            ­1.5___________________________________________________________________________Fold­change describes the relative expression of the given factor in osteoblast cells treated for48 h with 35 nM PDE7 or PDE8A siRNAs.

55

TABLE 8. Summary of qPCR results. Effects of PDE7 and PDE8A RNAi silencing on geneexpression in hMSC­derived human osteoblasts.Gene PDE7 silencing PDE8A silencing

OCN ↑ ↔β­catenin ↑ ↔Caspase 8 ↑ ↓CREB­5 ↑ ↔VDR ↓ ↔BMP­1 ↔ ↓BMP­5 ↑ ↔PDE7A ↓ ↔PDE7B ↓ ↔PDE8A ↔ ↓PDE4A ↔ ↔PDE4B ↔ ↔Ostoglycin ↔ ↓

The effects of gene expression are indicated as follows: up­regulation ↑, down­regulation ↓,no effect ↔.

The  BMPs  are  known  to  be  important  in  the  process  of  osteoblast  differentiation,proliferation,  apoptosis  and  morphogenesis  (Canalis  et  al.  2003).  The  BMP  family  can  bedivided into several subgroups based on BMP gene sequence identity. BMP­1 is unrelated tothe other BMPs and does not regulate the growth and differentiation of skeletal cells (Uzel etal.  2001).  BMP­5,  BMP­6,  BMP­7  and  BMP­8  belong  to  the  same  subgroup  and  all  haveosteoinductive properties  (Wozney 2002). In our study, PDE7  inhibition caused a threefoldup­regulation of BMP­5 gene expression. The BMPs have been shown to interact closely withWnt  peptides.  The  Wnt  proteins  are  secreted  by  osteoblasts  and  are  required  for  BMP­mediated stimulation of osteoblast proliferation and function (Nakashima et al. 2004, Nelsonand Nusse 2004). Inhibition of PDE7 up­regulated  ­catenin, a protein that plays a central rolein  the  Wnt  signalling  pathway  (Rawadi  and  Roman­Roman  2005),  suggesting  that  PDE7could  participate  in  Wnt­mediated  effects  on  osteoblast  differentiation.  In  parallel  withdifferentiation,  osteoblasts  show  increased  apoptosis,  which  is  established  to  be  associatedwith the activation of caspases (Canalis et al. 2005). Evidence indicates that caspase­mediatedactivation of apoptosis  is  necessary  for osteoblastic differentiation  (Mogi and Togari 2003).The  effect  of  PDE7  silencing  was  to  significantly  up­regulate  caspase­8.  In  contrast  tocaspase­8,  vitamin  D  receptor  (VDR)  mRNA  was  reduced  in  PDE7­silenced  cells.  VDRknock­out  mice  show  increased  bone  volume  and  density    compared  with  wild­type  mice(Tanaka  and  Seino  2004),  and  osteoblasts  lacking  VDR  have  enhanced  differentiation,manifested as an increase in both ALP activity and  mineralized matrix formation (Sooy et al.2005). The  increased expression of caspase­8 and the decreased expression of VDR seen  inPDE7­silenced  cells  therefore  support  PDE7  inhibition  promoting  an  osteoblastic  gene­expression pattern. The transcription factors of the CREB family stimulate basal transcription

56

of  CRE­containing  genes  and  mediate  induction  of  transcription  upon  phosphorylation  byprotein  kinase  A.  CREB  has  been  demonstrated  to  have  a  major  role  in  the  expression  ofskeleton­specific genes (Sakamoto et al. 1998). In our study, silencing of PDE7 up­regulatedCREB­5, which further suggests that PDE7 is involved in gene expression during osteoblasticdifferentiation.

In  human  T  cells,  PDE7  has  the  potential  to  regulate  cell  function,  including  cytokineproduction,  proliferation and expression of activation  markers  (Nakata et al.  2002). PDE8Asilencing was also performed, resulting in decreased BMP­1, osteoglycin and collagen 1 geneexpression. As the expression of these genes is known to be associated with the osteoblasticphenotype, PDE8A inhibition seems unlikely to enhance the differentiation of osteoblasts. Infact,  PDE8  silencing  appears  to  have  a  negative  effect  on  osteoblastic  differentiation.  Themicroarray  findings  of  PDE7  silencing  illustrate  up­regulation  of  several  osteogenic  geneexpressions, which could reflect augmented osteoblastic differentiation.

5.5 PDE7 inhibition enhances cAMP signalling, differentiation and mineralization in hMSC­derived osteoblasts (II)

The  effects  of  PDE7  silencing on  cAMP  accumulation  were  studied  to  evaluate  the  role  ofPDE7 on cAMP signalling. The cells were cultured  in differentiation  media  for  two weeks.PDE7 was silenced with siRNAs for 48 h, and cAMP synthesis was then stimulated with 10µM forskolin for 20 min. PDE7 silencing increased the forskolin­stimulated cAMP responsemore  than  threefold.  As  the  PDE7  silencing  increased  both  cAMP  accumulation  andosteocalcin  expression,  we  investigated  the possible  involvement  of  PKA  in  PDE7  siRNA­stimulated  osteocalcin  expression.  Treatment  with  5  µM  PKA  inhibitor  H­89  after  siRNAtransfections completely abolished the effects of PDE7 siRNAs on OCN expression.

To  evaluate  bALP  activity  in  hMSC­derived  osteoblasts,  the  transfections  were  performedafter  13  days  of  differentiation  and  bALP  activity  was  measured  two  weeks  aftertransfections.  The  treatment  with  PDE7  siRNAs  significantly  increased  the  bALP  levelscompared  with  control  cells.  Silencing  PDE8  did  not  significantly  change  bALP.  Inundifferentiated hMSC cells, bALP levels were low (Figure 8A). Mineralization was assessedby  Alizarin  red  S  staining.  Cells  were  treated  with  siRNAs  or  with  20  µM  of  the  PDE7selective  inhibitor  BRL  50481  (3­(N,N­dimethylsulfonamido)­4­methyl­nitrobenzene).  After21­22 days of differentiation, the mineralization of PDE7 siRNA­treated cells was increasedthreefold compared with controls (Figure 8B). Treatment with PDE8A siRNAs had no effecton  mineralization  (Figure  8C).  Treatment  with  20  µM  BRL­50481  also  increasedmineralization (Figure 8D).

57

Control

siRNAPDE7

H­89

siRNAPDE7 +

 H­89

0

100

200 **

*** *** % C

hang

e

Control

5 nM

10 nM

15 nM

35 nM

100 n

M0

1000

2000

3000

4000

**** **

µM

/mg 

prot

ein

Control

5 nM

10 nM

15 nM

35 nM

100 n

M0

250

500

750

µM

/mg 

prot

ein

Control

20 µM B

RL0

250

500

750

1000***

µM

/mg 

prot

ein

(A) (B)

(C) (D)

Figure  8.  Effect  of  PDE7  and  PDE8A  silencing  with  siRNAs  on  ALP  activity  andmineralization  in  osteoblasts  differentiated  from  hMSC  cells.  (A)  To  evaluate  the  bALPactivity, the cells were grown on 96­well plates for two weeks in differentiation media. Aftera  two­week  treatment  with  PDE7­  or  PDE8A­specific  siRNAs,  bone­specific  ALP  activity(bALP)  was  assayed  by  immunoassay.    To  determinate  the  effect  of  PDE7  and  PDE8A  onmineralization, the siRNAs  transfections were performed after 12­14 days of differentiation,and mineralization was assayed by Alizarin red S staining after 21­22 days of differentiation.Treatment  with  PDE7  siRNAs  dose­dependently  increased  mineralization  (B),  whereasPDE8 siRNA treatment did  not  (C). Treating  the cultures with  the PDE7 selective  inhibitorBRL­50481 (20 µM)  for 21 days also  increased mineralization compared with controls (D).Results represent means ±SEM of triplicate determinations (**P<0.01,**P<0.001).

58

Silencing  of  PDE7  increased  both  bALP  activity  and  mineralization  in  hMSC­derivedosteoblasts,  implying  that  PDE7  also  has  an  impact on  osteoblast  cell  functions.  Our  studyshows  that  inhibition  by  PDE7­specific  siRNAs  and  by  the  chemical PDE7  inhibitor  BRL­50481 can control osteoblastic differentiation and mineralization.

5.6 Down­regulation of cAMP­PDE by dexamethasone in human osteosarcoma cells (III)

The effect of glucocorticoids on the function of PDEs was studied at the mRNA and enzymelevels  in vitro.  Dex,  a  synthetic  glucocorticoid,  has  been  shown  to  increase  cell  osteogenicdifferentiation and proliferation in bone marrow mesenchymal stem cell (Bellows et al. 1987,Shalhoub et al. 1992, Jaiswal et al. 1997, Bellows et al. 1998), but impair the proliferative andmetabolic activity of osteoblasts (Patschan et al. 2001). Treatment of MG­63 cells with Dexsignificantly decreased  the cAMP­PDE activity  in a dose­ and  time­dependent  manner. Theeffect  of  Dex  on  cAMP­PDE  activity  was  reduced  by  up  to  50%,  with  the  half­maximalinhibition of  the cAMP­PDE activity being 5 nM. Treatment with 100 nm Dex for 48 h didnot  affect  the  cGMP­PDE  activity  in  MG­63  cells,  indicating  that  Dex  specifically  affectscAMP­PDE activity. The effect of Dex on the cAMP­PDE activity of SaOS­2 cells was lesspronounced,  probably  due  to  the  lower  PDE4  activity  in  these  cells.  The  sensitivity  ofhomogenates  towards  PDE  inhibitors  EHNA  and  cilostamide  was  not  affected  by  Dextreatment. However, homogenates of the Dex­treated cells were, as expected, less sensitive torolipram. The Dex treatment also resulted in homogenates that were slightly less sensitive to8­MMX. Stimulation with forskolin or rolipram separately only increased the cAMP levels inthe cells moderately. Simultaneous treatment with rolipram and forskolin resulted in 20­foldand  60­fold  increases  in  cAMP  levels  in  MG­63  and  SaOS­2  cells,  respectively.  Treatmentwith  Dex  did  not  increase  cAMP  levels  in  either  cell  type.  However,  the  Dex  treatmentsignificantly  reduced  the  potency  of  rolipram  to  increase  forskolin­stimulated  cAMPresponse.  We  detected  a  substantial  decrease  in  the  cAMP­specific  PDE  activity  aftertreatment with Dex. The Dex treatment alone did not affect the intracellular cAMP levels, buttaking  into account  that PDE4  is abundant PDE type  in  both MG­63 and SaOS­2 cells,  thereduction of the cAMP response in the Dex­treated cell might better describe the effect of analtered  PDE  profile  of  the  cells.  A  previous  study with  SaOS­2  cells  has presented  similarresults, showing that Dex alone does not affect cAMP levels (Mori et al. 1999).

5.7  Influence  of  dexamethasone  on  PDE  subtype  mRNA  expression  in  MG­63  and  SaOS­2cells (III)

Expression of the following cAMP­PDE subtypes was detected by reverse transcriptase PCR:PDE1A,  PDE1C,  PDE2A,  PDE3A,  PDE4A,  PDE4B,  PDE4C,  PDE4D,  PDE7A,  PDE7B,PDE8A,  PDE10A  and  PDE11A  (Table  5).  Of  these  subtypes,  Dex  reduced  the  mRNA  ofPDE4A and PDE4B subtypes in MG­63 and SaOS­2 cells. The decrease in PDE4 mRNA was

59

dose­  and  time­dependent,  with  the  half­maximal  inhibition  being  5–10 nM  Dex,  withsignificant effects  seen within 4 h of  treatment. Further analysis  revealed a  reduction  in  themRNA of  the PDE4A4 and PDE4B1  isoforms. Transcription of PDE4A1, 4A7, 4A10, 4B2and 4B3 was also evident, but we could not detect any reduction with Dex treatment in theseisoforms.  The  observed  effects  on  cAMP­PDE  activity  and  cAMP  accumulation  wereaccompanied  by  a  decrease  in  the  mRNA  levels  of  PDE4A4  and  PDE4B1  isoforms.  Thesepresented  results  thereby  support  the  hypothesis  that  PDEs  play  a  role  in  Dex­mediatedeffects  on  cAMP  signalling  in  osteoblasts.  In  rat  cultured  osteoblasts,  two  glucocorticoidactions  have  been  shown:  the  stimulation  of  adenylate  cyclase  and  the  inhibition  of  PDE(Chen  and  Feldman  1979).    In  addition,  treatment  with  Dex  increases  PTHR  mRNAexpression  in  ROS  17/2.8  cells,  but  not  in  opossum  kidney  cells.  Glucocorticoids  mightenhance  sensitivity  to  PTH  by  changing  the  number  of  PTH  receptors  and  their  affinity  toPTH; this suggests the involvement of a post­transcriptional mechanism (Urena et al. 1994).

5.8 Effects of PDEs in osteoblasts

In summary, Studies I­III revealed that the profiles of PDEs in human osteoblasts are similarto these in rat and mouse osteoblasts. The main PDEs are PDE1, PDE2, PDE3, PDE4, PDE7and  PDE8,  but  families  and  subtypes  can  vary  depending  on  differentiation  stage  andcircumstances around the cells. In Study II, we investigated the effects of silencing of PDE7and PDE8 on gene expression and differentiation in hMSC­derived osteoblasts. Inhibition ofPDE7 enhances differentiation and mineralization. The inhibitors of PDE4 have earlier beenshown to increase osteoblastic differentiation and bone formation. At the sequence level, thePDE7  family  is  closely  related  to  the  PDE4  family  (Omori  et  al.  2007).  This  could  be  onereason  why  effects  of  PDE  inhibition  are  osteogenic  in  hMSC­derived  osteoblasts.  Asynthetic glucocorticoid, Dex, could alter  the profile of PDE4 subtypes  in osteoblastic cells.Dex did not have a direct effect on  intracellular  cAMP  levels,  but did cause  a  reduction  incAMP response. Our results suggest an additional mechanism by which glucocorticoids couldbe involved in bone loss.

Only a small number of studies exist on the effects of PDEs in bone or bone cells, and most ofthese have  investigated  the effect of PDE4  inhibitors.  In  rats  inoculated with  Walker 256/Scarcinoma to induce an artificial osteoporotic state, bone loss was reduced by over 50% aftera two­week treatment with the PDE4 selective inhibitor denbufylline (Miyamoto et al. 1997).PDE4  inhibitors  rolipram,  XT­44  and  XT­611  have  been  shown  to  increase  mineralizednodule  formation  (Waki  et  al.  1999)  and  PGE2­mediated  anabolic  effects  (Miyamoto  et  al.2003)  in  rat  bone  marrow  culture.  XT­44  also  increases  the  bone  mineral  density  ofovariectomized  rats.  Rolipram  and  pentoxifylline,  a  non­selective  inhibitor,  have  beendemonstrated  to  increase  both  cortical  and  cancellous  bone  mass  in  mice  (Kinoshita  et  al.2000).  Daily  injections  of  pentoxifylline  have  been  reported  to  enhance  bone  formation  inBMP­impregnated  collagen  disks  implanted  into  the  back  muscles  of  mice  (Horiuchi  et  al.

60

2001).  Pentoxifylline  also  promotes  the  osteoblastic  differentiation  of  ST2  mouse  bonemarrow­derived  stromal  cells,  although  a  recent  report  suggests  that  the  effects  ofpentoxifylline are protein kinase A­independent (Rawadi et al. 2001). More information aboutthe  role  of  other  PDEs  in  bone  formation  and  how  these  enzymes  are  involved  in  theregulation of bone metabolism is needed.

Future studies should clarify the profile of PDEs in various differentiation stages,  from stemcells to osteoblasts, osteocytes and osteoclast. The role of PDE7 in bone metabolism could beinvestigated  in  vivo  with,  for  instance,  knock­out  mice.  T­cell  functional  studies  haverevealed that PDE7A knock­out mice have  approximately 10% lower body fat than PDE7A ­/+  and  PDE7A  ­/­  mice  (Yang  et  al.  2003).  The  lower  body  fat  in  PDE7A  knock­out  micemight also indicate changes in bone tissue.

5.9 Influences of bioactive peptides IPP, VPP and LKP on gene expression and proliferationin hMSC­derived osteoblasts (IV)

The  Hum­16K  cDNA  microarray  data  analysis  revealed  that  the  bioactive  peptide  IPPregulated more genes in hMSC­derived osteoblasts than the bioactive peptides VPP and LKP.Treatments  with  20  µM  tripeptides  showed  that  IPP  up­regulated  270  genes  and  down­regulated 100 genes. The respective numbers  for VPP were 25 and 10, and  for LKP 16 and14. IPP not only regulated a larger number of genes than VPP or LKP, but also was the onlytripeptide to up­regulate osteogenic differentiation factors.

The  PCA  included  245/270  up­regulated  genes  and  47/100  down­regulated  genes  in  thedifferent components. The PCA revealed 45% of the variation in up­regulated genes from theIPP­treated  osteoblasts  to  be  attributed  to  the  first  principal  component  (up­regulated  celldifferentiation). The second principal component explained 25.3% (up­regulated cell growth)and  the  third  22.6%  (up­regulated  cell  transcription)  of  the  variation  (Table  9).  The  firstcomponent (115 genes) had heavy positive accumulations of genes related to differentiation,receptors,  mitochondria,  apoptosis,  enzymes,  signal  transduction,  membrane  architecture,membrane  function  and  membrane  transport. The  second  component  (47  genes)  had  heavypositive  accumulations  of  genes  related  to  cell  adhesion,  osteogenic  differentiation  and  cellgrowth  and  proliferation,    while  and  third  component  (83  genes)  had  heavy  positiveassociations  with  genes  related  to  transcription,  splicing,  translation,  EST  and  ‘unknownfunctions’.  The  up­regulation  of  these  genes  indicates  that  IPP  enhances  osteoblastproliferation and differentiation.

Our  results  also  show  that  IPP  increases  UMR­106  and  hMSC  proliferation,  but  notproliferation  of  mature  osteoblasts.  When  the  state  of  differentiation  increases,  the

61

TABLE  9. Summary  of  the  microarray  analysis:  number  of  up­regulated  (Up)  and  down­regulated  (Down)  genes  by  IPP,  VPP  and  LKP  inosteoblasts differentiated from hMSCs and the average fold change.

EST, expressed sequence tags.

62

proliferation of the mature osteoblast become limited (Amedee et al. 1994). These results arein  line  with  ours.  Biomaterial  studies  conducted  with  an  Arg­Gly­Asp­containing  peptidesequence  also  reveal  increased  primary  calvarial  osteoblast  cell  proliferation  in  response  tothe peptide  (Huag  et  al.  2003,  Benoit  and  Anseth 2005).  A  set  of  IPP­regulated genes wasfurther  analysed  by  qRT­PCR  (Figure  9).  Microarray  analysis  had  revealed  IPP­treatedosteoblasts to express more differentiation genes (BMP­5), transcription factor genes (CREB­5)  and  hormone­related  genes  (PTHrP)  than  control  cells.  Apoptosis­related  caspase­8  andvitamin D receptor gene expression, were decreased due to IPP treatment. In this study, 24­htreatment with IPP upregulated PTHrP and CREB­5 genes. PTH exerts anabolic and cataboliceffects  in  vivo  on  bone  (Dempster  et  al.  1993).  In  vitro  studies on  osteoblasts  have  showncontinuous treatment with very low concentration of PTH to stimulate UMR­106­01 and

Figure  9.  The  gene  expression  of  IPP  treated  osteoblasts  analysed  by  a  cDNA  microarraymethod ( ) and quantitative real­time PCR (qRT­PCR; ). At  least 1.8­fold difference wasused  in  control v.  treatment  comparison  in  microarray  data,  and  the  ratio  in  the  untreatedsample was used as a control (100). qRT­PCR results were normalized by amplification of theresult  of  ­actin.  Values  are  means,  with  their  standard  errors  depicted  by  vertical  bars,  oftriplicate  determinations,  where  all  amplified  genes  have  their  own  controls.  Mean  valueswere significantly different from those of the controls (unpaired Student's t test): ***P<0.001,**P<0.01,*P<0.05.

63

primary  osteoblast  proliferation  (Dempster  et  al.  1993,  Swarthout  et  al.  2001).  PTHrp  is  apolypeptide hormone sharing a common sequence with PTH (Walsh et al. 1997), and both canfunction  with  the PTHR  (Walsh  et  al.  1997).  PTHrp  produced by  osteoblasts  is  thought  tofunction  locally.  Within  the  skeletal  microenvironment,  PTHrp  propels  pluripotential  bonemarrow  stromal  cells  towards  the  osteogenic  lineage  and  also  exerts  antiapoptotic  effects(Karaplis  and  Goltzman  2000).  The  ability  of  PTH  to  drive  changes  in  gene  expression  isdependent upon activation of such transcription factors as the CREB (Swarthout et al. 2002).In  response  to  PTH,  CREB  is  phosphorylated  and  in  turn  activates  transcription.  Theincreased  proliferation  in  IPP­treated  hMSC  could  be  due  to  increased  PTHrp  and  CREB.Because  cell  proliferation  and  maturation  are  sequential  processes,  increased  numbers  ofpluripotent  cells  and  their  augmented  maturation  into  osteoblasts  could  increase  the  totalnumber of osteoblasts. The ability of IPP to induce BMP­5 production can increase osteoblastdifferentiation  and  augment  fracture  repair,  a  feature  associated  with  a  protein­rich  diet(Bastow et al. 1983, Delmi et al. 1990, Rizzoli and Bonjour 2004).

In  summary,  IPP,  one  of  the  three  investigated  tripeptides,  increase  the  gene  expression  ofosteogenic  factors  in  hMSC­derived  osteoblasts.  IPP  has  also  shown  to  increase  matrixproduction, mineralization and decrease RANKL/OPG ration  in  the same cells  (Huttunen etal. 2007). Tripeptides could transport via hPepT1 transporter both in the small intestine and inthe  proximal  tubule  or  hPepT2  in  the  proximal  tubule  (see  Nielsen  and  Brodin  2003,Steffansen  et  al.  2004),  but  the  transporter  and  signaling  mechanisms  in  bone  cells  is  stillunclear.  Forthcoming  studies  could  focus  on  the  the  receptor  interactions  and  cellularmechanisms of bioactive tripeptides in bone cells.

6. Conclusions

We  identified  PDE  families  and  their  subtypes  present  in  normal  human  osteoblasts  and  inhuman osteosarcoma (SaOS­2, MG­63) cell  lines. The main difference  in PDE profiles wasthe obvious  lack of PDE3 activity  in osteosarcoma cells, while PDE3 contributed  largely tothe total PDE activity of normal human osteoblasts. The major PDE subtype of MG­63 cellswas PDE4, and MG­63 cells also contained higher levels of PDE4 activity than SaOS­2 cells.The reason for this difference in PDE profiles is unknown. The osteosarcoma cell lines mightrepresent  a  different  differentiation  stage  that  is  associated  with  high  levels  of  PDE4expression, or PDE profiles might have altered during cell transformation from the normal tothe malignant phenotype.

Our  results  showed  that  inhibition  by  PDE7­specific  siRNAs  and  by  the  chemical  PDE7inhibitor BRL­50481 can control osteoblastic differentiation and mineralization. The findingsalso  indicated  that gene­specific silencing with the RNAi method is useful  for  inhibiting thegene  expression  of  specific  PDEs.  The  silencing  of  PDE7  up­regulates  several  osteogenic

64

genes at the mRNA level, and PDE7 may therefore play an important role in the regulation ofosteoblastic differentiation.

At the mRNA and enzyme levels, Dex, a synthetic glucocorticoid, decreased gene expressionof PDE4  subtypes 4A and 4B,  but did  not affect other cAMP­PDEs  in MG­63 and SaOS­2cells.  The  observed  effects  on  cAMP­PDE  activity  and  cAMP  accumulation  wereaccompanied  by  a  decrease  in  the  mRNA  levels  of  PDE4A4  and  PDE4B1  isoforms.  Theresults  support  the  hypothesis  that  PDEs  play  a  role  in  Dex­mediated  effects  in  cAMPsignalling  in  osteoblasts.  PDEs  appears  to  be  involved  in  a  mechanism  by  whichglucocorticoids affect cAMP signalling, demonstrating a new possible route in the formationof glucocorticoid­induced bone loss involving the down­regulation of cAMP­PDE.

Based  on  our  results,  tripeptide  IPP  increases  cell  proliferation  at  the  cellular  and  mRNAlevels. Genes associated with differentiation were up­regulated and apoptosis  inducers weredown­regulated. These findings indicate that IPP enhances gene expression in such a way thatincreases the differentiation of hMSC into osteoblasts and prolongs their viability.

This work has  shown that PDEs possess mechanisms that enhance osteoblast differentiationand affect cAMP signalling by glucocorticoids. Bioactive peptide IPP augmented the potentialto up­regulate osteogenic and cell viability genes  in osteoblasts. Bioactive peptides, such asIPP,  might  well  contribute  to  the  positive  effects  that  dietary  protein  has  on  bone  mineralmass.  Understanding  the  role  of  PDEs  in  bone  cells  is  important  since  bone  metabolism,especially  bone  formation,  could  be  affected  by  PDEs.  These  enzymes  are  potentiallysignificant targets for the development of new signal transduction therapies  for the treatmentof metabolic bone diseases, for example osteoporosis. However, additional research is neededto elucidate the mechanisms of PDEs and bioactive peptides on osteoblast differentiation andbone formation.

65

7. Acknowledgements

This work was carried out at the Calcium Research Unit, Division of Nutrition, Department ofApplied  Chemistry  and  Microbiology,  University  of  Helsinki.  I  am  grateful  to  ProfessorsAntti Ahlström and Marja Mutanen for providing excellent working facilities.

I also wish to thank Professor Tapio Palva, Head of the Division of Genetics, Department ofEnvironmental  and  Biosciences,  for  interest  in  my  work  and  for  guiding  me  through  thebureaucracy.

My  warmest  gratitude  is  due  to  my  supervisors,  Dr.  Christel  Lamberg­Allardt  for  theopportunity to work with her group and Dr. Mikael Ahlström for introducing me to the worldof  bone  cells  and  phosphodiesterases.  Both  (of  you)  gave  me  a  great  deal  of  scientificfreedom, guidance when needed and were always ready to listen to my ideas.

I  am  grateful  to  the  reviewers,  Dr.  Petri  Lehenkari  and  Dr.  Anitta  Mahonen,  their  advice,constractive criticism and insightful comments that substantially improved this thesis.

I am grateful  to my author­editor, Carol Ann Pelli,  for revising the English  language of themanuscript.

I  thank  present  and  past  colleagues  at  the  Calcium  Research  Unit;  Heini  Karp,  MerjaKärkkäinen, Virpi Kemi, Marika Laaksonen, Zahirul Islam, Heli Viljakainen and particularlyMinna  Huttunen  for  stimulating  teamwork,  help  and  enthusiasm,  and  Ulrike  Riehle  for  hercontribution  and  for  being  such  a  nice  person.  It  has  been  a  pleasure  to  share  the  ups  anddowns of academic life with all of you.

My heartfelt  thanks are due to my  friends; providing  for relaxing company and all kinds offun  over  the  years.  I  greatly  appreciate  the  unwavering  support  of  my  mother  Kerttu,  mybrother Harri, my mother­in­law Elise and her husband Juhani. Finally, my deepest love andgratitude are owed to my husband Pasi and dauhgter Iisa. I am so  lucky to have you  in  mylife.

This work has been financially supported by the Finnish Graduate School in Musculo­SkeletalDiseases,  the  National  Technology  Agency  of  Finland  (TEKES),  the  Oskar  ÖflundFoundation, the Finnish Konkordia Fund, the Finnish Cultural Foundation, the Ella ja GeorgEhrnrooth Foundation and the Betty Väänänen Foundation.

Sipoo, March 2008

Minna Pekkinen

66

8. References

Abe, E., Yamamoto, M., Taguchi, Y., Lecka­Czernik, B., O’Brien, C., Economides, A.N.,Stahl, N., Jilka, R.L., Manolagas, S.C. (2000) Essential requirement of BMPs­2/4 for bothosteoblast  and  osteoclast  formation  in  murine  bone  marrow  cultures  from  adult  mice:antagonism by noggin. J Bone Miner Res 15:663­673.Advani,  S.,  LaFrancis,  D., Bogdanovic, E.,  Taxel,  P.,  RaisZ,  L.G.; Kream,  B.E.  (1997)Dexamethasone suppresses  in vivo levels of bone collagen synthesis  in neonatal mice. Bone20, 41­46.Ahlström, M., Lamberg­Allardt, C. (1997) Rapid protein kinase A­mediated activation ofcyclic AMP­phosphodiesterase by parathyroid  hormone  in UMR­106 osteoblast­like cells.  JBone Miner Res 12, 172­178.Ahlström,  M.,  Lamberg­Allardt,  C.  (1999)  Regulation  of  Adenosine  3`,5´­cyclicmonophosphate  (cAMP)  accumulation  in  UMR­106  osteoblast­like  cells:  Role  of  cAMP­phoshodiesterase and cAMP efflux. Bio Pharmacol 58, 1335­1340.Ahlström,  M.,  Lamberg­Allardt,  C.  (2000)  Inactivation  of  atrial  natriuretic  factor­stimulated cyclic guanosine 3`,5´­monophosphate (cGMP) in UMR­106 osteoblast­like cells.Bio Pharmacol 59, 1133­1139.Amedee, J., Bareille, R., Rouais, F., Cunningham, N., Reddi, H., Harmand, M.F. (1994)Osteogenin (bone morphogenic protein 3) inhibits proliferation and stimulates differentiationof osteoprogenitors in human bone marrow. Differentiation 58, 157­164.Anthony,  T.L.,  Brooks,  H.L.,  Boassa,  D.,  Leonov,  S.,  Yanochko,  G.M.,  Regan,  J.W.,Yool,  A.J.  (2000)  Cloned  human  aquaporin­1  is  a  cyclic  GMP­gated  ion  channel.  MolPharmacol 57, 576­88.Aravind, L.,  Koonin,  E.V.  (1998)  The  HD  domain  defines  a  new  superfamily  of  metal­dependent phosphohydrolases. Trends Biochem Sci 23, 469­472.Aubin, J.E. and Liu, F. (1996) The osteoblast  lineage. In: Bilezikian, J.P., Raisz, L.G. andRodan, G.A., Editors, 1996. Principles of Bone Biology, Academic Press, New York, pp. 51–67.Banerjee, C., McCabe, L.R., Choi, J­Y., Hiebert, S.W., Stein, J.L., Stein, G.S., Lian, J.B.(1997)  Runt  homology  domain  proteins  in  osteoblast  differentiation:  AML­3/CBFA1  is  amajor component of a bone specific complex. J Cell Biochem 66, 1­8.Bantounas,  I., Phylactou, L.A., Uney, J.B.  (2004) RNA  interference and  the use of  smallinterfering  RNA to study gene  function  in  mammalian  systems. J Mol Endocrinol 33, 545­457.Bastow, M.D.,  Rawlings,  J.,  Allison,  S.P.  (1983)  Benefits  of  supplementary  tube  feedingafter  fractured neck of  femur:  a  randomised controlled  trial. Br Med  J Clin Res 287, 1589­1592.Beavo, J.A. (1995) Cyclic nucleotide phosphodiesterases: Functional implications of multipleisoforms. Physil Rev 75, 725­748.Beazely, M.A., Watts, V.J. (2006) Regulatory properties of adenylate cyclases type 5 and 6:A progress report. European Journal of Pharmacology 535, 1­12.

67

Beard, M.B., Olsen, A.E., Jones, R.E., Erdogan, S., Houslay, M.D., Bolger, G.B.  (2000)UCR1  and  UCR2  domains  unique  to  the  cAMP­specific  phosphodiesterase  family  form  adiscrete module via electrostatic interactions. J Biol Chem 275, 10349­58.Bellows, C.G.,  Aubin,  J.E.,  Heersche,  J.N.  (1987) Physiological  concentrations  ofglucocorticoids  stimulate  formation of bone nodules  from  isolated rat calvaria cells  in vitro.Endocrinology, 121, 1985­1992.Bellows, C.G., Ciaccia, A., Heersche, J.N. (1998) Osteoprogenitor cells  in cell populationsderived  from  mouse  and  rat  calvaria differ  in  their  response  to  corticosterone,  cortisol,  andcortisone. Bone 23, 119­125.Bender, A.T., Ostenson, C.L., Wang, E.H., Beavo, J.A. (2005) Selective up­regulation ofPDE1B2 upon monocyte­to­macrophage differentiation. Proc Natl Acad Sci USA 102, 497­502.Benoit, D.S., Anseth, K.S. (2005) The effect on osteoblast function of colocalized RGD andPHSRN epitopes on PEG surfaces. Biomaterials 26, 5209­5220.Benovic,  J.L.,  Strasser,  R.H.,  Caron,  M.G,  Lefkowitz,  R.J.  (1986)  Beta­adrenergicreceptor  kinase:  identification  of  a  novel  protein  kinase  that  phosphorylates  the  agonist­occupied form of the receptor. Proc Natl Acad Sci U S A 83, 2797­801.Bian, J.S., Zhang, W.M., Pei,  J.M., Wong, T.M. (2000) The  role of phosphodiesterase  inmediating  the  effect  of  protein  kinase  C  on  cyclic  AMP  accumulation  upon  kappa­opioidreceptor stimulation in the rat heart. J Pharmacol Exp Ther 292, 1065­1070.Birnbaumer, L.  (1990) G proteins  in  signal  transduction.  Ann Rev Pharmacol Toxicol 30,675­705.Bloom,  T.J.,  Beavo,  J.A.  (1996)  Identification  and  tissue­specific  expression  of  PDE7phosphodiesterase splice variants. Proc Natl Acad Sci U S A 93, 14188­92.Boden, S.D., Hair, G., Titus, L., Racine, M., McCuaig, K., Wozney, J.M., Nanes, M.S.(1997) Glucocorticoid­induced differentiation of fetal rat calvarial osteoblasts is mediated bybone morphogenetic protein­6. Endocrinology 138:2820­2828.Boerth,  N.J.,  Dey,  N.B.,  Cornwell,  T.L.,  Lincoln,  T.M.  (1997)  Cyclic  GMP­dependentprotein kinase regulates vascular smooth muscle cell phenotype. J Vasc Res 34, 245­59.Bolger,  G.,  Michaeli,  T.,  Martins,  T.,  St  John,  T.,  Steiner,  B.,  Rodgers,  L.,  Riggs,  M.,Wigler, M., Ferguson, K. (1993) A family of human phosphodiesterases homologous to thedunce learning and memory gene product of Drosophila melanogaster are potential targets forantidepressant drugs. Mol Cell Biol 13, 6558­71.Bolger,  G.B.,  Erdogan,  S.,  Jones,  R.E.,  Loughney,  K.,  Scotland,  G.,  Hoffmann,  R.,Wilkinson, I., Farrell, C., Houslay, M.D. (1997) Characterization of five different proteinsproduced by alternatively spliced mRNAs from the human cAMP­specific phosphodiesterasePDE4D gene. Biochem J 328, 539­48.Borton, A.J., Frederick, J.P., Datto, M.B., Wang, X.F., Weinstein, R.S. (2001) The loss ofSmad3  results  in  a  lower  rate  of  bone  formation  and  osteopenia  through  dysregulation  ofosteoblast differentiation and apoptosis. J Bone Mineral Res 16, 1754­64.

68

Boyce, B.F., Xing, L., Jilka, R.J., Bellido, T., Weinstein, R.S., Parfitt, A.M., Manolagas,S.C.  (2002)  Apoptosis  in  bone  cells.  In:  Bilezikian,  J.P.,  Raisz,  L.G.,  Rodan,  G.A.  (eds.)Principles of Bone Biology, 2nd  ed. Academic Press, San Diego, CA, USA, pp 151­168.Boguslawski,  G.,  Hale,  L.V.,  Yu,  X­P.,  Miles,  R.R.,  Onyia,  J.E.,  Santerre,  R.F.,Chandrasekhar,  S.  (2000)  Activation  of  Osteocalcin  Transcription  Involves  Interaction  ofProtein Kinase A­ and Protein Kinase C­dependent Pathways. J Biol Chem 275, 999­1006.Butcher, R.W., Sutherland, E.W. (1962) Adenosine 3',5'­phosphate in biological materials.I.  Purification  and  properties  of  cyclic  3',5'­nucleotide  phosphodiesterase  and  use  of  thisenzyme to characterize adenosine 3',5'­phosphate in human urine. J Biol Chem 237, 1244­50.Calvi, L.M., Sims, N.A., Hunzelman, J.L., Knight, M.C., Giovannetti, A., Saxton, J.M.,Kronenberg,  H.M.,  Baron,  R.,  Schipani,  E.  (2001)  Activated  parathyroidhormone/parathyroid  hormone­related  protein  receptor  in  osteoblastic  cells  differentiallyaffects cortical and trabecular bone. J Clin Invest 107, 277­86.Canalis, E. (2005) Mechanisms of glucocorticoid action  in bone. Curr Osteopor Rep 3, 98­102.Canalis, E., Bilezikian, J.P., Angeli, A., Giustina, A. (2004) Perspectives on glucocorticoid­induced osteoporosis. Bone 34, 593­598.Canalis, E., Deregowski, V., Pereira, R.C., Gazzerro, E. (2005) Signals that determine thefate of osteoblastic cells. J Endocrinol Invest 28, 3­7.Canalis,  E.,  Economides,  A.N.,  Gazzerro,  E.  (2003)  Bone  morphogenetic  proteins,  theirantagonists, and the skeleton. Endocr Rev 24, 218­235.Canalis, E.,  Giustina,  A.,  Bilezikian,  J.P. (2007) Mechanisms  of  anabolic  therapies  forosteoporosis. New Engl J Med 357, 905­916.Chabance, B.,  Marteau,  P.,  Rambaud,  J.C.,  Migliore­Samour,  D.,  Boynard,  M.,Perrotin, P., Guillet, R., Jolles, P., Fiat, A.M. (1998) Casein peptide release and passage tothe blood in humans during digestion of milk or yogurt. Biochimie 80, 155­165.Chasin,  M.,  Harris,  D.N.  (1976)  Inhibitory  and  activators  of  cyclic  nucleotidephosphodiesterase. Adv Cycl Nucleotide Res 7, 225­64.Chen, T.L., Feldman, D.  (1979) Glucocorticoid  receptors and actions  in  subpopulations ofcultured  rat bone cells.  Mechanism of dexamethasone potentiation of parathyroid  hormone­stimulated cyclic AMP production. J Clin Invest 63, 750­758.Cheng,  J.,  Thompson,  M.A.,  Walker,  H.J.,  Gray,  C.E.,  Diaz  Encarnacion,  M.M.,Warner, G.M., Grande, J.P. (2004) Differential regulation of mesangial cell mitogenesis bycAMP phosphodiesterase isozymes 3 and 4. Am J Physiol Renal Physiol 287, F940­953.Choi,  J­Y.,  Pratap,  J.,  Javed,  A.,  Zaidi, S.K., van Wijnen,  A.J.,  Lian,  J.B.,  Stein,  J.L.,Jones, S.C., Stein, G.S. (2000)  In vivo replacement of  the Runx/Cbfa1 gene with a mutantgene  lacking a subnuclear targeting signal and transcriptional regulatory  functions results  insevere skeletal abnormalities. J Bone Miner Res 15, S157.Church, VL., Francis­West, P.  (2002) Wnt signalling during limb development.  Int J DevBiol 46 (7),927­36.

69

Collins, S.P., Uhler, M.D. (1999) Cyclic AMP­ and cyclic GMP­dependent protein kinasesdiffer in their regulation of cyclic AMP response element­dependent gene transcription. J BiolChem 274, 8391­404.Conti,  M.,  Iona,  S.,  Cuomo,  M.,  Swinnen,  J.V.,  Odeh,  J.,  Svoboda,  M.E.  (1995)Characterization of a hormone­inducible, high affinity adenosine 3'­5'­cyclic monophosphatephosphodiesterase from the rat Sertoli cell. Biochemistry 34, 7979­87.Conti, M., Jin, S.L.C. (2000) The molecular biology of cyclic nucleotide phoshodiesterases.Prog Nuleic Acid Res Mol Biol 63, 1­38.Conti, M., Richter, W., Mehats, C., Livera, G., Park, JY., Jin, C. (2003) Cyclic AMP­specificPDE4  phosphodiesterases  as  critical  components  of  cyclic  AMP  signaling. J  Biol  Chem.278:5493­6.Corbin,  J.D.,  Turko,  I.V.,  Beasley,  A.,  Francis,  S.H.  (2000)  Phosphorylation  ofphosphodiesterase­5  by  cyclic  nucleotide­dependent  protein  kinase  alters  its  catalytic  andallosteric cGMP­binding activities. Eur J Biochem 267, 2760­7.Cote, R.H. (2004) Characteristics of photoreceptor PDE (PDE6): similarities and differencesto PDE5.  Int J Impot Res 16 Suppl 1, S28­33.Coudray, C., Charon, C., Komas, N., Mory, G., Diot­Dupuy, F., Manganiello, V., Ferre,P.,  Bazin,  R.  (1999)  Evidence  for  the  presence  of  several  phosphodiesterase  isoforms  inbrown adipose  tissue of Zucker  rats: modulation  of PDE2 by  the  fa gene expression. FEBSLett  456, 207­10.Craig, W.S. Cheng, S., Mullen, D.G., Blevitt, J., Pierschbacher, M.D. (1995) Concept andprogress  in  the  development  of  RGD­containing  peptide  pharmaceuticals.  Biopolymers  37,157­175.Dalton,  G.D.,  Dewey,  W.L.  (2006)  Protein  kinase  inhibitor  peptide  (PKI):  a  family  ofendogenous neuropeptides that modulate neuronal cAMP­dependent protein kinase function.Neuropeptides 40, 23­34.Daub, H., Weiss, F.U., Wallasch, C., Ullrich, A. (1996) Role of transactivation of the EGFreceptor in signaling by G­protein coupled receptors. Nature 379, 557­560.Dee,  K.C.  (1996a)  Considerations  for  design  of  proactive  dental/orthopaedic  implantbiomaterials. Ph.D. Thesis, Rensselaer Polytechnic Institute, Troy, New York, 1996.Dee, K.C.,  Rueger,  D.C.,  Andersen,  T.T.,  Bizios,  R.  (1996b)  Conditions  which  promotemineralization  at  the  bone­implant  interface:  a  model  in  vitro  study.  Biomaterials  17,  209­215.Dee, K.C.,  Andersen,  T.T.,  Bizios,  R.  (1998)  Design  and  function  of  novel  osteoblast­adhesive peptides  for  chemical  modification  of  biomaterials.  J Biomed Mater  Res  40,  371­377.Degerman, E.,  Smith,  C.J.,  Tornqvist,  H.,  Vasta,  V.,  Belfrage,  P.,  Manganiello,  V.C.(1990)  Evidence  that  insulin  and  isoprenaline  activate  the  cGMP­inhibited  low­Km  cAMPphosphodiesterase in rat fat cells by phosphorylation. Proc Natl Acad Sci U S A 87, 533­7.Degerman,  E.,  Belfrage,  P.,  Manganiello,  V.C.  (1997)  Structure,  localization,  andregulation of cGMP­inhibited phosphodiesterase (PDE3). J Biol Chem 272, 6823­6.

70

Delay,  R.J.,  Dubin,  A.E.,  Dionne,  V.E.  (1997)  A  cyclic  nucleotide­dependent  chlorideconductance in olfactory receptor neurons. J Membr Biol 159, 53­60.Delmi, M.,  Rapin,  C.H.,  Bengoa,  J.M.,  Delmas,  P.D.,  Vasey,  H.,  Bonjour,  J.P.  (1990)Dietary  supplementation  in  elderly  patients  with  fractured  neck  of  the  femur. Lancet  335,1013­1016.Dempster,  D.W.,  Cosman,  F.,  Parisien,  M.,  Shen,  V.  (1993)  Anabolic  actions  ofparathyroid hormone on bone. Endocr Rev 14, 690­709.Dessauer,  C.W.,  Gilman,  A.G.  (1997) The  catalytic  mechanism  of  mammalian  adenylylcyclase.  Equilibrium  binding  and  kinetic  analysis  of  P­site  inhibition.  J  Biol  Chem  272,27787–27795.Dodge­Kafka,  K.L.,  Soughayer,  J.,  Pare,  G.C.,  Carlisle  Michel,  J.J.,  Langeberg,  L.K.,Kapiloff,  M.S.,  Scott,  J.D.  (2005)  The  protein  kinase  A  anchoring  protein  mAKAPcoordinates two integrated cAMP effectors pathways. Nature 22, 574­578.Dortch­Carnes, J., Potter, D.E. (2002) Inhibition of cAMP accumulation by kappa­receptoractivation  in  isolated  iris­ciliary  bodies:  role  of  phosphodiesterase  and  protein  kinase  C.  JPharmacol Exp Ther 301, 599­604.Drissi, H., Hott, M., Marie, P.J.,   Lasmoles, F.  (1997) Expression of  the CT/CGRP geneand its regulation by dibutyryl cyclic adenosine monophosphate in human osteoblastic cells. JBone Mineral Res 12, 1805­1814.Ducy,  P.,  Zhang,  R.,  Geoffrey,  V.,  Ridall,  A.L.,  Karsenty  G.  (1997) Osf2/Cbfa1  :  Atranscriptional activator of osteoblast differentiation. Cell 89, 747­754.Ducy, P., Schinke, T., Karsenty G. (2000) The osteoblast: A sophisticated fibroblast undercentral surveillance.Science 289, 1501­1504.Dulak, J., Jozkowicz, A., Dembinska­Kiec, A., Guevara, I., Zdzienicka, A., Zmudzinska­Grochot, D., Florek, I., Wojtowicz, A., Szuba, A., Cooke, J.P. (2000) Nitric oxide inducesthe synthesis of vascular endothelial growth factor by rat vascular smooth muscle cells.Arterioscler Thromb Vasc Biol 20, 659­666.Eastell,  R.,Reid,DM., Compston,  J., Cooper, C., Fogelman,  I., Francis, RM., Hosking, DJ.,Purdie, DW., Ralston, SH., Reeve, J., Russell, RG., Stevenson, JC., Torgerson, DJ. (1998) AUK  Consensus  Group  on  management  of  glucocorticoid­induced  osteoporosis:  an  update.Journal of Internal Medicine 244(4):271­92.Ehses,  J.A.,  Casilla,  V.R.,  Doty,  T.,  Pospisilik,  J.A.,  Winter,  K.D.,  Demuth,  H.U.,Pederson,  R.A.,  McIntosh,  C.H.  (2003) Glucose­dependent  insulinotropic  polypeptidepromotes beta­(INS­1) cell survival via cyclic adenosine monophosphate­mediated caspase­3inhibition and regulation of p38 mitogen­activated protein kinase.  Endocrinology 144, 4433­45.Eigenthaler,  M.,  Lohmann,  S.M.,  Walter,  U.,  Pilz,  RB.  (1999)  Signal  transduction  bycGMP­dependent protein kinases and their emerging roles  in the regulation of cell  adhesionand gene expression. Rev Physiol Biochem Pharmacol 135, 173­209.Elbashir, S.M., Harborth, J., Weber, K., Tuschl, T.  (2002)  Analysis of gene  function  insomatic mammalian cells using small interfering RNAs. Methods (Duluth) 26, 199­213.

71

el­Husseini, A.E., Bladen, C., Vincent, S.R. (1995) Molecular characterization of a  type IIcyclic GMP­dependent protein kinase expressed in the rat brain. J Neurochem 64, 2814­7.Engels,  P.,  Fichtel,  K.,  Lubbert,  H.  (1994)  Expression  and  regulation  of  human  and  ratphosphodiesterase type IV isogenes. FEBS Lett 350, 291­5.Engels, P., Sullivan, M., Muller, T., Lubbert, H. (1995) Molecular cloning and functionalexpression in yeast of a human cAMP­specific phosphodiesterase subtype (PDE IV­C). FEBSLett 358, 305­10.Engleman, V.W.,  Nickols,  G.A.,  Ross,  F.P.,  Horton,  M.A.,  Griggs,  D.W.,  Settle,  S.L.,Ruminski, P.G., Teitelbaum, S.L. (1997) A peptidomimetic antagonist of the alpha(v)beta3integrin inhibits bone resorption in vitro and prevents osteoporosis in vivo. J Clin Invest 99,2284­2292.Erba, D., Ciappellano, S., Testolin, G. (2002) Effect of the ratio of casein phosphopeptidesto calcium (w/w) on passive calcium transport  in the distal small  intestine of rats. Nutrition18,  743­746.Farndale, R.W., Sandy, J.R., Atkinson, S.J., Pennington, S.R., Meghji, S., Meikle, M.C.(1988)  Parathyroid  hormone  and  prostaglandin  E2  stimulate  both  inositol  phosphates  andcyclic AMP accumulation in mouse osteoblast cultures. Biochem J 252, 263­268.Favot,  L.,  Keravis,  T.,  Lugnier,  C.  (2004)  Modulation  of  VEGF­induced  endothelial  cellcycle  protein  expression  through  cyclic  AMP  hydrolysis  by  PDE2  and  PDE4.  ThrombHaemost  92,  634­45.Fawcett,  L.,  Baxendale,  R.,  Stacey,  P.,  McGrouther,  C.,  Harrow,  I.,  Soderling,  S.,Hetman,  J.,  Beavo,  J.A.,  Pillips,  S.C.  (2000)  Molecular  cloning  and  characterization  of  adistinct human phoshodiesterase gene family: PDE 11A. Proc Natl Acad Sci USA 97, 3702­3707.Fidock,  M., Miller,  M.,  Lanfear,  J.  (2002)  Isolation  and  differential  tissue  distribution  oftwo human cDNAs encoding PDE1 splice variants. Cell Signal 14, 53­60.Finkelstein,  J.S., Klibanski, A.,  Schaefer,  E.H., Hornstein, M.D.,  Schiff,  I.,  Neer,  R.M.(1994) Parathyroid hormone for the prevention of bone loss  induced by estrogen deficiency.N Engl J Med 331, 1618­1623.Fischer, O.M., Hart, S., Gschwind, A., Ullrich, A. (2003) EGFR signal  transactivation  incancer cells. Biochem Soc Trans 31, 1203­1208.Fisher,  D.A.,  Smith,  J.F.,  Pillar,  J.S.,  St  Denis,  S.H.,  Cheng,  J.B.  (1998)  Isolation  andcharacterization  of  PDE8A,  a  novel  human  cAMP­specific  phosphodiesterase.  BiochemBiophys Res Commun 246, 570­7.Fisher, D.A.,  Smith,  J.F.,  Pillar,  J.S.,  St  Denis,  S.H.,  Cheng,  J.B.  (1998)  Isolation  andcharacterization of PDE9A, a  novel human cGMP­specific phosphodiesterase.  J Biol Chem273, 15559­64.Forte,  L.R.,  Currie,  M.G.  (1995)  Guanylin:  a  peptide  regulator  of  epithelial  transport.FASEB J 9, 643­50.Foulkes, N.S., Borrelli, E., Sassone­Corsi, P. (1991) CREM gene: use of alternative DNA­binding domains generates multiple antagonists of cAMP­induced transcription. Cell 64, 739­749.

72

Francis, S.H., Colbran, J.L., McAllister­Lucas, L.M. Corbin, J.D. (1994) Zinc interactionsand conserved motifs of the cGMP­binding cGMP­specific phosphodiesterase suggest that itis a zinc hydrolase. J Biol Chem 269, 22477­80.Francis, S.H., Corbin, J.D. (1994) Progress in understanding the mechanism and function ofcyclic GMP­dependent protein kinase. Adv Pharmacol 26, 115­70.Francis, S.H., Bessay, E.P., Kotera, J., Grimes, K.A., Liu, L., Thompson, W.J., Corbin,J.D.  (2002)  Phosphorylation  of isolated  human  phosphodiesterase­5  regulatory  domaininduces  an  apparent  conformational  change  and  increases  cGMP  binding  affinity.  J  BiolChem 277, 47581­7.Frandsen,  E.K.,  Krishna,  G.  A.  (1976) A  simple  ultrasensitive  method  for  the  assay  ofcyclic AMP and cyclic GMP in tissues. Life Sci 76,529­541.French,  P.J.,  Bijman,  J.,  Edixhoven,  M.,  Vaandrager,  A.B.,  Scholte,  B.J.,  Lohmann,S.M.,  Nairn,  A.C.,  de  Jonge,  H.R.  (1995)  Isotype­specific  activation  of  cystic  fibrosistransmembrane conductance  regulator­chloride channels by cGMP­dependent protein kinaseII. J Biol Chem 270, 26626­31.Fitzgerald,  R.J.,  Murray,  B.A.  (2006)  Bioactive  peptides  and  lactic  fermantations.  Int  JDairy Technol 59, 118­125.Fu, Q., Jilka, R.L., Manolagas, S.C., O'Brien, C.A. (2002) Parathyroid hormone stimulatesreceptor  activator  of  NFkappa  B  ligand  and  inhibits  osteoprotegerin  expression  via  proteinkinase  A  activation  of  cAMP­response  element­binding  protein.  J  Biol  Chem  277,  48868­48875.Fuglsang, A.,  Rattray,  F.P.,  Nilsson,  D.,  Nyborg,  N.C.B.  (2003)  Lactic  acid  bacteria:inhibition of angiotensin converting enzyme in vitro and in vivo. Anton Leeuw Int J G 83, 27­34.Fujishige,  K.,  Kotera,  J.,  Michibata,  H.,  Yuasa,  K.,  Takebayashi,  S.,  Okumura,  K.,Omori,  K.  (1999) Cloning  and  characterization  of  a  novel  human  phosphodiesterase  thathydrolyzes both cAMP and cGMP (PDE10A). J Biol Chem 274, 18438­45.Fujishige, K., Kotera, J., Omori, K. (1999) Striatum­ and testis­specific phosphodiesterasePDE10A isolation and characterization of a rat PDE10A. Eur J Biochem 266, 1118­27.Fujishige,  K.,  Kotera,  J.,  Yuasa,  K.,  Omori,  K.  (2000)  The  human  phosphodiesterasePDE10A gene genomic organization and evolutionary relatedness with other PDEs containingGAF domains. Eur J Biochem 267, 5943­51.Gamanuma, M.,  Yuasa,  K.,  Sasaki,  T.,  Sakurai,  N.,  Kotera,  J.,  Omori,  K.  (2003)Comparison  of  enzymatic  characterization  and  gene  organization  of  cyclic  nucleotidephosphodiesterase 8 family in humans. Cell Signal 15, 565­74.Gambaryan, S., Hausler, C., Markert, T., Pohler, D., Jarchau, T., Walter, U., Haase, W.,Kurtz, A., Lohmann, S.M. (1996) Expression of type II cGMP­dependent protein kinase inrat kidney  is  regulated by dehydration and correlated with renin gene expression. J Clin Inv98, 662­70.Gardner, C., Robas, N., Cawkill, D., Fidock, M. (2000) Cloning and characterization of thehuman  and  mouse  PDE7B,  a  novel  cAMP­specific  cyclic  nucleotide  phosphodiesterase.Biochem Bioph Res 272, 186­92.

73

Geoffroy,  V., Fouque,  F.,  Nivet,  V.,  Clot, J.P.,  Lugnier,  C., Desbuquois,  B.,  Benelli C.(1999)  Activation of a cGMP­stimulated cAMP phosphodiesterase by protein kinase C  in aliver Golgi­endosomal fraction. Eur J Biochem 259, 892­900.Geoffroy, V., Fouque, F., Lugnier, C., Desbuquois, B., Benelli, C. (2001) Characterizationof an in vivo hormonally regulated phosphodiesterase 3 (PDE3) associated with a liver Golgi­endosomal fraction. Arch Biochem Biophys 387, 154­62.Gilman, A.G. (1990) Regulation of adenylyl cyclase by G proteins. Adv Second MessengerPhosphoprotein Res 24,51–57.Giorgi, M.,  Giordano,  D.,  Rosati,  J.,  Tata,  A.M.,  Augusti­Tocco,  G.  (2002)  Differentialexpression and localization of CaM­dependent phosphodiesterase genes during ontogenesis ofchick dorsal root ganglion. J Neurochem 80, 970­9.Glavas, N.A., Ostenson, C., Schaefer, J.B., Vasta, V., Beavo, J.A. (2001) T cell activationup­regulates cyclic nucleotide phosphodiesterases 8A1 and 7A3. Proc Natl Acad Sci U S  A98, 6319­24.Conti,  M.,  Beavo,J.  (2007) Biochemistry  and  physiology  of  cyclic  nucleotidephosphodiesterases: essential components in cyclic nucleotide signaling. Annu Rev Biochem76, 481­511.Goraya, T.A., Masada, N., Ciruela, A., Cooper, D.M.  (2004) Sustained entry of Ca2+  isrequired  to  activate  Ca2+­calmodulin­dependent  phosphodiesterase  1A.  J  Biol  Chem  279,40494­504.Gori, F., Hofbauer, L.C., Dunstan, C.R., Spelsberg, T.C., Khosla, S., Riggs, B.L. (2000)The expression of osteoprotegerin and RANK ligand and the support of osteoclast formationby  stromal­osteoblast  lineage  cells  is  developmentally  regulated.  Endocrinology  141,  4768­76.Gordeladze,  J.O.,  Drevon,  C.A.,  Syversen,  U.,  Reseland,  J.E.  (2002)  Leptin  stimulateshuman osteoblastic cell proliferation, de novo collagen synthesis, and mineralization: Impacton differentiation markers, apoptosis, and osteoclastic signaling. J Cell Biochem 85, 825­836.Grange,  M.,  Sette,  C.,  Cuomo,  M.,  Conti,  M.,  Lagarde,  M.,  Prigent,  A.F.,  Nemoz,  G.(2000)  The  cAMP­specific  phosphodiesterase  PDE4D3  is  regulated  by  phosphatidic  acidbinding.  Consequences  for  cAMP signaling pathway  and  characterization  of  a  phosphatidicacid binding site. J Biol Chem 275, 33379­87.Gross­Langenhoff, M., Hofbauer, K., Weber, J., Schultz, A., Schultz, J.E. (2006) cAMPis a  ligand  for  the  tandem  GAF domain of  human phosphodiesterase 10 and cGMP  for  thetandem GAF domain of phosphodiesterase 11. J Biol Chem 281, 2841­6.Gudi,  T.,  Huvar,  I.,  Meinecke,  M.,  Lohmann,  S.M.,  Boss,  G.R.,  Pilz,  R.B.  (1996)Regulation of gene expression by cGMP­dependent protein kinase. Transactivation of the c­fos promoter. J Biol Chem 271, 4597­600.Gudi, T., Lohmann, S.M., Pilz, R.B. (1997) Regulation of gene expression by cyclic GMP­dependent  protein  kinase  requires  nuclear  translocation  of  the  kinase:  identification  of  anuclear localization signal. Mol Cell Biol 17, 5244­54.Guipponi, M., Scott, H.S., Kudoh, J., Kawasaki, K., Shibuya, K., Shintani, A., Asakawa,S., Chen, H., Lalioti, M.D., Rossier, C., Minoshima, S., Shimizu, N., Antonarakis, S.E.

74

(1998) Identification and characterization of a novel cyclic nucleotide phosphodiesterase gene(PDE9A) that maps to 21q22.3: alternative splicing of mRNA transcripts, genomic structureand sequence. Hum Genet 103, 386­92.Haass, N.K.,  Smalley,  K.S.,  Li,  L.,  Herlyn,  M.  (2005)  Adhesion,  migration  andcommunication in melanocytes and melanoma. Pigment Cell Res 18, 150­159.Haby,  C.,  Lisovoski,  F.,  Aunis,  D.,  Zwiller,  J.  (1994)  Stimulation  of  the  cyclic  GMPpathway  by  NO  induces  expression  of  the  immediate  early  genes  c­fos  and  junB  in  PC12cells. J Neurochem 62, 496­501.Hadjidakis, D.J., Androulakis, I.I. (2006) Bone remodeling. Ann NY Acad Sci  1092, 385­396.Hagen, V., Dzeja, C., Bendig, J., Baeger, I., Kaupp, U.B. (1998) Novel caged compoundsof  hydrolysis­resistant  8­Br­cAMP  and  8­Br­cGMP:  photolabile  NPE  esters.  J  PhotochemPhotobiol B 42, 71­8.Halene, T.B., Siegel, S.J. (2007) PDE inhibitors in psychiatry­­future options for dementia,depression and schizophrenia? Drug Discov Today 12, 870­878.Han, P., Zhu, X., Michaeli, T. (1997) Alternative splicing of the high affinity cAMP­specificphosphodiesterase  (PDE7A)  mRNA  in  human  skeletal  muscle and  heart.  J Biol  Chem 272,16152­7.Hannon, G.J. (2002) RNA interference. Nature 418, 244­251.Hata, Y., Yamamoto, M., Ohni, M., Nakajima, K., Nakamura, Y., Takano, T. (1996) Aplacebo­controlled study of the effect of sour milk on blood pressure in hypertensive subjects.Am J Clin Nutr 64, 767­771.Hebb,  A.L.,  Robertson,  H.A.,  Denovan­Wright,  E.M.  (2004)  Striatal  phosphodiesterasemRNA  and  protein  levels  are  reduced  in  Huntington's  disease  transgenic  mice  prior  to  theonset of motor symptoms. Neuroscience 123, 967­81.Heino, T.J., Hentunen, T.A., Vaananen, H.K.  (2002) Osteocytes  inhibit osteoclastic  boneresorption  through  transforming  growth  factor­beta:  enhancement  by  estrogen.  J  CellBiochem 85, 185­197.Heino, T.J., Hentunen, T.A., Vaananen, H.K. (2004) Conditioned medium from osteocytesstimulates the proliferation of bone marrow mesenchymal stem cells and their differentiationinto osteoblasts. Exp Cell Res 294, 458­468.Hentunen,  T.A.,  Lakkakorpi,  P.T.,  Tuukkanen,  J.,  Lehenkari,  P.P.,  Sampath,  T.K.,Vaananen,  H.K.  (1995)  Effects  of  recombinant  human  osteogenic  protein­1  on  thedifferentiation  of  osteoclast­like  cells  and  bone  resorption. Biochem  Biophys  Res  Commun209, 433­443.Hernandez­Ledesma, B., Amigo, L., Ramos, M., Recio, I. (2004) Angiotensin convertingenzyme  inhibitory  activity  in  commercial  fermented  products.  Formation  of  peptides undersimulated gastrointestinal digestion. J Agric Food Chem 52, 1504­1510.Hetman, J.M.,  Soderling,  S.H.,  Glavas,  N.A.,  Beavo,  J.A.  (2000)  Cloning  andcharacterization  of  PDE7B,  a  cAMP­specific  phosphodiesterase.  Proc  Natl  Acad  Sci  USA97,472­6.

75

High, W.B. (1987) Effects of orally administered prostaglandin E­2 on cortical bone turnoverin adult dogs: a histomorphometric study. Bone 8, 363­373.Hodsman,  AB.,  Bauer,  DC.,  Dempster,  DW.,  Dian,  L.,  Hanley,  DA.,  Harris,  ST.,Kendler, DL., McClung, MR., Miller, PD., Olszynski, WP., Orwoll, E., Yuen, CK. (2005)Parathyroid  hormone  and  teriparatide  for  the  treatment  of  osteoporosis:  a  review  of  theevidence and suggested guidelines for its use. Endocrine Reviews 2005;26(5):688­703.Hofbauer,  L.C.,  Gori,  F.,  Riggs,  B.L.,  Lacey,  D.L.,  Dunstan,  C.R.,  Spelsberg,  T.C.,Khosla,  S.  (1999)  Stimulation  of  osteoprotegerin  ligand  and  inhibition  of  osteoprotegerinproduction  by  glucocorticoids  in  human  osteoblastic  lineage  cells:  potential  paracrinemechanisms of glucocorticoid­induced osteoporosis. Endocrinology 140, 4382­4389.Hoffmann, R., Baillie, G.S., MacKenzie, S.J., Yarwood, S.J., Houslay, M.D. (1999) TheMAP  kinase  ERK2  inhibits  the  cyclic  AMP­specific  phosphodiesterase  HSPDE4D3  byphosphorylating it at Ser579. EMBO J 18, 893­903.Hori, G., Wang, M.F., Chan, Y.C., Komatsu, T., Wong, Y., Chen, T.H., Yamamoto, K.,Nagaoka,  S.,  Yamamoto,  S.  (2001)  Soy  protein  hydrolyzate  with  bound  phospholipidsreduces  serum  cholesterol  levels  in  hypercholesterolemic  adult  male  volunteers.  BiosciBiotechnol Biochem 65, 72­78.Horiguchi, N., Horiguchi, H., Suzuki, Y. (2005) Effect of wheat gluten hydrolysate on theimmune system in healthy human subjects. Biosci Biotechnol Biochem 69, 2445­2449.Houslay, M.D. (2001) PDE4 cAMP­specific phosphodiesterases. Prog Nucleic Acid Res MolBiol 69, 249­315.Houslay, MD., Adams, DR. (2003)  PDE4 cAMP phosphodiesterases: modular enzymes thatorchestrate  signalling  cross­talk,  desensitization  and  compartmentalization.  BiochemicalJournal. 370:1­18.Hsu, H., Lacey, D.L., Dunstan, C.R., Solovyev, I., Colombero, A., Timms, E., Tan, H.L.,Elliott, G., Kelley, M.J., Sarosi, I., Wang, L., Xia, X.Z., Elliott, R., Chiu, L., Black, T.,Scully,  S.,  Capparelli,  C.,  Morony,  S.,  Shimamoto,  G.,  Bass,  M.B.,  Boyle,  W.J.  (1999)Tumor necrosis factor receptor family member RANK mediates osteoclast differentiation andactivation induced by osteoprotegerin ligand. Proc Natl Acad Sci U S A 96, 3540­3545.Hu, H., McCaw, E.A., Hebb, A.L., Gomez, G.T., Denovan­Wright, E.M. (2004) Mutanthuntingtin affects the rate of transcription of striatum­specific  isoforms of phosphodiesterase10A. Eur J Neurosci 20, 3351­63.Huag,  H.,  Zhao,  Y.,  Liu,  Z.,  Zhang,  Y.,  Zhang,  H.,  Fu,  T.,  Ma,  X.  (2003)  Enhancedosteoblast functions on RGD immobilized surface. J Oral Implantol 29, 73­79.Huang,  L.J.,  Durick,  K.,  Weiner,  J.A.,  Chun,  J.,  Taylor,  S.S.  (1997)  Identification  of  anovel  protein  kinase  A  anchoring  protein  that  binds  both  type  I  and  type  II  regulatorysubunits. J Biol Chem 272, 8057­64.Huang,  Z.J.,  Edery,  I.,  Rosbash,  M.  (1993)  PAS  is  a  dimerization  domain  common  toDrosophila period and several transcription factors. Nature 364, 259­62.Hughes,  F.J.,  Collyer,  J.,  Stanfield,  M.,  Goodman,  S.A. (1995)  The  effects  of  bonemorphogenetic  protein­2,  ­4,  and  ­6  on  differentiation  of  rat  osteoblast  cells in  vitro.Endocrinology 136:2671­2677.

76

Huttunen, M.M., Pekkinen, M., Ahlström, M.E.B, Lamberg­Allardt, C.J.E. (2007) Long­term  effects  of  tripeptide  Ile­Pro­Pro on  osteoblast  differentiation  in  vitro.  J  Nutr  Biochemdoi:10.1016/j.jnutbio.2007.09.006.Idriss, S.D., Gudi, T., Casteel, D.E., Kharitonov, V.G., Pilz, R.B., Boss, G.R. (1999) Nitricoxide  regulation  of  gene  transcription  via  soluble  guanylate  cyclase  and  type  I  cGMP­dependent protein kinase. J Biol Chem 274, 9489­93.Ishizuya, T., Yokose, S., Hori, M., Noda, T., Suda, T., Yoshiki, S., Yamaguchi, A. (1997)Parathyroid  hormone  exerts  disparate  effects  on  osteoblast  differentiation  depending  onexposure time in rat osteoblastic cells. J Clin Invest 99, 2961­2970.Itoh,  K.,  Udagawa,  N.,  Katagiri,  T.,  Iemura,  S.,  Ueno,  N.,  Yasuda,  H.,  Higashio,  K.,Quinn,  J.M.,  Gillespie,  M.T.,  Martin,  T.J.,  Suda,  T.,  Takahashi,  N.  (2001) Bonemorphogenetic  protein  2  stimulates  osteoclast  differentiation  and  survival  supported  byreceptor activator of nuclear factor­kappaB ligand. Endocrinology 142,  3656­3662.Iwata,  K.,  Li,  J.,  Follet,  H.,  Phipps,  R.J.,  Burr,  D.B.  (2006)  Bisphosphonates  suppressperiosteal  osteoblast  activity  independently  of  resorption  in  rat  femur  and  tibia. Bone  39,1053­8.Jaiswal, N.,  Haynesworth,  S.E.,  Caplan,  A.I.,  Bruder,  S.P.  (1997)  Osteogenicdifferentiation of purified,  culture­expanded  human  mesenchymal  stem cells  in  vitro. J CellBiochem 64, 295­312.James, I.E.,  Dodds,  R.A.,  Olivera,  D.L.,  Nuttall,  M.E.,  Gowen,  M.  (1996)  Humanosteoclastoma­derived stromal cells: correlation of the ability to form mineralized nodules invitro with formation of bone in vivo. J Bone Miner Res 11, 1453­1460.Jee, W.S., Ueno, K., Deng, Y.P., Woodbury, D.M. (1985) The effects of prostaglandin E2in  growing  rats:  increased  metaphyseal  hard  tissue  and  cortico­endosteal  bone  formation.Calcif Tissue Int 37, 148­157.Jee, W.S., Ueno, K., Kimmel, D.B., Woodbury, D.M., Price, P., Woodbury, L.A. (1987)The role of  bone cells  in  increasing metaphyseal hard  tissue  in  rapidly growing  rats  treatedwith prostaglandin E2. Bone 8, 171­178.Jeon, M.J., Kim, J.A., Kwon, S.H., Kim, S.W., Park, K.S., Park, S.W., Kim, S.Y., Shin, C.S.(2003)  Activation  of  peroxisome  proliferator­activated  receptor­gamma  inhibits  the  Runx2­mediated transcription of osteocalcin in osteoblasts. J Biol Chem 278, 23270­7.Jiang, Y., Vaessen,  B.,  Lenvik,  T.,  Blackstad,    M., Reyes,  M.,  Verfaillie,  C.M.  (2002a)Multipotent progenitor cells can be isolated from postnatal murine bone marrow, muscle, andbrain, Exp Hematol 30, 896–904.Jiang,  Y.,  Jahagirdar,  B.N.,  Reinhardt,  R.L.,  Schwartz,  R.E.,  Keene,  C.D.,  Ortiz­Gonzalez,  X.R.,  Reyes,  M.,  Lenvik,  T.,  Lund,  T.,  Blackstad,  M.,  Du,  J.,  Aldrich,  S.,Lisberg,  A.,  Low,  W.,  Largaespada,  D.A.,  Verfaillie,  C.M.  (2002b)  Pluripotency  ofmesenchymal stem cells derived from adult marrow, Nature 418, 41–49.Jiang, Y., Henderson, D., Blackstad, M., Chen, A., Miller, R.F., Verfaillie, C.M. (2003)Neuroectodermal  differentiation  from  mouse  multipotent  adult  progenitor  cells.  Proc  NatlAcad Sci USA 100 (Suppl. 1), 118554–118560.

77

Jilka, R.L.  (2007) Molecular and cellular mechanisms of the  anabolic effect of  intermittentPTH. Bone 40, 1434­1446.Jilka, R.L., Weinstein, R.S., Bellido, T., Parfitt, A.M., Manolagas, S.C. (1998) Osteoblastprogrammed  cell  death  (apoptosis):  modulation  by  growth  factors  and  cytokines.  J  BoneMiner Res 13, 793­802.Jilka,  R.L.,  Weinstein,  R.S.,  Bellido,  T.,  Roberson,  P.,  Parfitt,  A.M.,  Manolagas,  S.C.(1999)  Increased  bone  formation  by  prevention  of  osteoblast  apoptosis  with  parathyroidhormone. J Clin Invest 104, 439­46.Karaplis, A.C., Goltzman, D. (2000) PTH and PTHrP effects on the skeleton. Rev EndocrMetab Disord 1, 331­341.Kawaguchi,  H.,  Pilbeam,  C.C.,  Raisz,  L.G.  (1994)  Anabolic  effects  of  3,3  ,5­trooodothyronine  and  triiodothyroacetic  acid  in  cultured  neonatal  mouse  parietal  bones.Endocrinology 135, 971­976.Kayser, H.,  Meisel,  H.  (1996)  Stimulation  of  human  peripheral  blood  lymphocytes  bybioactive peptides derived from bovine milk proteins. FEBS Lett 383, 18­20.Keravis,  T., Komas,  N., Lugnier,  C.  (2000)  Cyclic  nucleotide  hydrolysis  in  bovine  aorticendothelial  cells  in  culture:  differential  regulation  in  cobblestone  and  spindle  phenotypes. JVasc Res 37, 235­49.Kerr, M.K., Churchill, G.A.  (2001) Experimental design  for gene expression  microarrays.Biostatistics 2:183­201.Khosla, S., Atkinson, E.J., Dunstan, C.R, O'Fallon, W.M. (2002) Effect of estrogen versustestosterone on circulating osteoprotegerin and other cytokine levels in normal elderly men.J Clin Endocrinol Metab  87, 1550­1554.Kinoshita, T. Kobayashi, S., Ebara, S., Yoshimura, Y., Horiuchi, H., Tsutsumimoto T.,Wakabayashi,  S.,  Takaoka,  K.  (2000)    Phosphodiesterase  inhibitors,  pentoxifylline  androlipram, increase bone mass mainly by promoting bone formation in normal mice. Bone 27,811­817.Kitamura,  T., Kitamura,  Y., Kuroda, S.,  Hino,  Y.,  Ando, M., Kotani, K., Konishi,  H.,Matsuzaki,  H.,  Kikkawa,  U.,  Ogawa,  W.,  Kasuga,  M.  (1999)  Insulin­inducedphosphorylation  and  activation  of  cyclic  nucleotide  phosphodiesterase  3B  by  the  serine­threonine kinase Akt. Mol Cell Biol 19, 6286­96.Kitts,  D.D.,  Weiler,  K.  (2003)  Bioactive  proteins  and  peptides  from  food  sources.Applications of bioprocesses used in isolation and recovery. Curr Pharm Des 9, 1309­1323.Kobayashi, T., Gamanuma, M., Sasaki, T., Yamashita, Y., Yuasa, K., Kotera, J., Omori,K.  (2003) Molecular comparison of rat cyclic nucleotide phosphodiesterase 8 family: uniqueexpression of PDE8B in rat brain. Gene 319, 21­31.Kogianni, G., Noble, B.S. (2007) The biology of osteocytes. Curr Osteoporos Rep 5, 81­6.Koh,  AJ.,  Beecher,  CA.,  Rosol,  TJ.,  McCauley,  LK.  (1999)  3',5'­Cyclic  adenosinemonophosphate  activation  in  osteoblastic  cells:  effects  on  parathyroid  hormone­1  receptorsand osteoblastic differentiation in vitro. Endocrinology. 140, 3154­62.

78

Komalavilas,  P.,  Lincoln,  T.M.  (1996)  Phosphorylation  of  the  inositol  1,4,5­trisphosphatereceptor.  Cyclic  GMP­dependent  protein  kinase  mediates  cAMP  and  cGMP  dependentphosphorylation in the intact rat aorta. J Biol Chem 271, 21933­8.Komori, T.  (2005) Regulation of  skeletal development by  the Runx  family of  transcriptionfactors. J Cell Biochem 95, 445­453.Komori, T.  (2006) Regulation of osteoblast differentiation  by  transcription  factors. . J CellBiochem 99, 1233­9.Konopleva,  M.,  Mikhail,  A.,  Estrov,  Z.,  Zhao,  S.,  Harris,  D.,  Sanchez­Williams,  G.,Kornblau,  S.M.,  Dong,  J.,  Kliche,  K.O.,  Jiang,  S.,  Snodgrass,  H.R.,  Estey,  E.H.,Andreeff, M. (1999) Expression and function of leptin receptor isoforms in myeloid leukemiaand myeloplastic syndromes: proliferative and anti­apoptotic activities. Blood 93, 1668­1676.Kopperud,  R.,  Krakstad,  C.,    Selheim,  F.,  Doskeland,  S.O.  (2003)  cAMP  effectormechanisms. Novel twists for an ‘old’ signaling system, FEBS Letters 546, 121–126.Korhonen, H.,  Pihlanto,  A.  (2003)  Food­derived  bioactive  peptides­­opportunities  fordesigning future foods. Curr Pharm Des 9, 1297­1308.Korhonen,  H.,  Pihlanto  A.  (2006)  Bioactive  peptides:  production  and  functionality.  IntDairy J 16, 945­960.Kulkarni,  NH.,  Halladay,  DL.,  Miles,  RR.,  Gilbert,  LM.,  Frolik,  CA.,  Galvin,  RJ.,Martin,  TJ.,  Gillespie,  MT.,  Onyia,  JE.  (2005)  Effects  of  parathyroid  hormone  on  Wntsignaling pathway in bone. Journal of Cellular Biochemistry. 95(6):1178­90.Kumegawa,  M.,  Ikeda,  E.,  Tanaka,  S.,  Haneji,  T.,  Yora,  T.,  Sakagishi,  Y.,  Minami,  N.,Hiramatsu,  M.  (1984)  The  effects  of  prostaglandin  E2,  parathyroid  hormone,  1,25dihydroxycholecalciferol,  and  cyclic  nucleotide  analogs  on  alkaline  phosphatase  activity  inosteoblastic cells. Calcif Tissue Int 36, 72­76.Kwak,  B.R.,  Saez,  J.C.,  Wilders,  R.,  Chanson,  M.,  Fishman,  G.I.,  Hertzberg,  E.L.,Spray, D.C., Jongsma, H.J. (1995) Effects of cGMP­dependent phosphorylation on rat andhuman connexin43 gap junction channels. Pflugers Archiv ­ European Journal of Physiology430, 770­8.Lalli,  E.,  Sassone­Corsi,  P.  (1994)  Signal  transduction  and  gene  regulation:  the  nuclearresponse to cAMP. J Biol Chem 269, 17359­62.Lanyon,  L.,  and  Skerry,  T.  (2001)  Postmenopausal  osteoporosis  as  a  failure  of  bone'sadaptation to functional loading: a hypothesis. J Bone Miner Res 16, 1937–1947.Lajeunesse, D., Frondoza, C., Schoffield, B., Sacktor, B. (1990) Osteocalcin secretion bythe human osteosarcoma cell line MG­63. J Bone Miner Res 5, 915­22.Lajeunesse, D., Kiebzak, G.M., Frondoza, C., Sacktor, B. (1991) Regulation of osteocalcinsecretion  by  human  primary  bone  cells  and  by  the  human  osteosarcoma  cell  line  MG­63.Bone Miner 14, 237–250.Lecka­Czernik,  B.,  Gubrik,  I.,  Moerman,  E.J.,  Kajkenova,  O.,  Lipschitz,  D.A.,Manolagas,  S.C.,  Jilka,  R.L.  (1999) Inhibition  of  Osf2/Cbfa1  expression  and  terminalosteoblast differentiation by PPARgamma2. J Cell Biochem 74, 357­371.Leroy,  M.J.,  Mehats,  C.,  Duc­Goiran,  P.,  Tanguy,  G.,  Robert,  B.,  Dallot,  E.,  Mignot,T.M.,  Grange,  G.,  Ferre,  F.  (1999)  Effect  of  pregnancy  on  PDE4  cAMP­specific

79

phosphodiesterase messenger ribonucleic acid expression in human myometrium. Cell Signal11,  31­37.Leskela, H.V., Risteli, J., Niskanen, S., Koivunen, J., Ivaska, K.K., Lehenkari, P. (2003)Osteoblast recruitment from stem cells does not decrease by age at  late adulthood. BiochemBiophys Res Commun 311, 1008­1013.Levitzki,  A.  (1997)  Targeting  signal  transduction  for  disease  therapy.  Medical  Oncology.14(2), 83­89.Lian, J.B, Stein, G.S., Aubin, J.E. Bone formation: Maturation and Functional Activities ofOsteoblast  Lineage  Cells.  In  Primers  on  the  Metabolic  Bone  Diseases  and  Disorders  ofMineral  Metabolism;  Favus,  M.J.,  Ed.;  American  Society  of  Bone  and  Mineral  Research:Washington D.C., 2003; Chapter 3.Lim,  J.,  Pahlke,  G.,  Conti, M.  (1999)  Activation  of  the  cAMP­specific  phosphodiesterasePDE4D3  by  phosphorylation.  Identification  and  function  of  an  inhibitory  domain.  J  BiolChem 274, 19677­85.Lincoln,  T.M., Dey, N.B., Boerth, N.J.,  Cornwell,  T.L.,  Soff,  G.A.  (1998)  Nitric  oxide­­cyclic  GMP  pathway  regulates  vascular  smooth  muscle  cell  phenotypic  modulation:implications in vascular diseases. Acta Physiol Scand 164, 507­15.Lincoln,  T.M.,  Wu,  X.,  Sellak,  H.,  Dey,  N.,  Choi,  C.S.  (2006)  Regulation  of  vascularsmooth  muscle  cell  phenotype  by  cyclic  GMP  and  cyclic  GMP­dependent  protein  kinase.Front Biosci 11, 356­367.Liu,  L.,  Underwood,  T.,  Li,  H.,  Pamukcu,  R.,  Thompson,  W.J.  (2002)  Specific  cGMPbinding by the cGMP binding domains of cGMP­binding cGMP specific phosphodiesterase.Cell Signal 14, 45­51.Logan, CY., Nusse, R. (2004) The Wnt signaling pathway in development and disease. AnnuRev Cell Dev Biol 20,781­810.Lohse, M.J., Benovic, J.L., Codina, J., Caron, M.G., Lefkowitz, R.J. (1990) beta­Arrestin:a protein that regulates beta­adrenergic receptor function. Science 248(4962):1547­50.Loughney,  K., Martins,  T.J., Harris,  E.A., Sadhu, K., Hicks,  J.B., Sonnenburg, W.K.,Beavo, J.A., Ferguson, K. (1996) Isolation and characterization of cDNAs corresponding totwo human calcium, calmodulin­regulated, 3',5'­cyclic nucleotide phosphodiesterases. J BiolChem 271, 796­806.Lugnier, C. (2006) Cyclic nucleotide phosphodiesterase (PDE) superfamily: a new target forthe development of specific therapeutic agents. Pharmacol Ther 109, 366­398.Luttrell, L., Daaka, Y., Lefkowitz, R. (1999) Regulation of tyrosine kinase cascades by G­protein­coupled receptors. Curr Opin Cell Biol 11, 177­183.Lykke­Andersen,  K.  (2006)  Regulation  of  gene  expression  in  mouse  embryos  and  itsembryonic cells through RNAi. Molecular Biotechnology. 34, 271­8.Lynch, M.J., Baillie, G.S., Mohamed, A., Li, X., Maisonneuve, C., Klussmann, E., vanHeeke,  G.,  Houslay,  M.D.  (2005)  RNA  silencing  identifies  PDE4D5  as  the  functionallyrelevant cAMP phosphodiesterase  interacting with beta arrestin to control  the protein kinaseA/AKAP79­mediated  switching  of  the  beta2­adrenergic  receptor  to  activation  of  ERK  inHEK293B2 cells. J Biol Chem 280, 33178­33189.

80

MacFarland,  R.T.,  Zelus,  B.D.,  Beavo,  J.A.  (1991) High  concentrations  of  a  cGMP­stimulated phosphodiesterase  mediate ANP­induced decreases  in cAMP and steroidogenesisin adrenal glomerulosa cells. J Biol Chem 266, 136­42.Mahonen, A., Pirskanen, A., Keinanen, R., Maenpaa, P.H. (1990) Effect of 1,25(OH)2D3on its receptor mRNA levels and osteocalcin synthesis in human osteosarcoma cells. BiochimBiophys Acta 1048, 30­37.Mahonen, A., Jukkola, A., Risteli, L., Risteli, J., Maenpaa, P.H. (1998) Type I procollagensynthesis  is  regulated by steroids and related hormones  in human osteosarcoma cells. J CellBiochem 68, 151­163.Majeska, R.J., Rodan, G.A. (1982) Alkaline phoshatase inhibition by parathyroid hormoneand  isoproterenol  in  a  clonal  rat osteosarcoma cell  line: Possible  mediation by cyclic  AMP.Calcif Tissue Int 34, 59­66.Manganiello, V.C., Murata, T., Taira, M., Belfrage, P., Degerman, E. (1995) Diversity incyclic nucleotide phoshodiesterase isozyme families. Arch Biochem Biophys 322, 1­13.Manning, C.D., McLaughlin, M.M., Livi, G.P., Cieslinski, L.B., Torphy, T.J., Barnette,M.S. (1996) Prolonged beta adrenoceptor stimulation up­regulates cAMP phosphodiesteraseactivity in human monocytes by increasing mRNA and protein for phosphodiesterases 4A and4B. J Pharmacol Exp Ther 276, 810­8.Martinez, S.E., Wu, A.Y., Glavas, N.A., Tang, X.B., Turley, S., Hol, W.G., Beavo, J.A.(2002) The two GAF domains in phosphodiesterase 2A have distinct roles in dimerization andin cGMP binding. Proc Natl Acad Sci USA 99, 13260­5.McCann, K.B.,  Shiell,  B.J.,  Michalski,  W.P.,  Lee,  A.,  Wan,  J.,  Roginski,  H.,  Coverty,M.J. (2006) Isolation and characterisation of a novel antibacterial peptide  from bovine  s1­casein. Int Dairy J 16, 316­323.Mehats, C.,  Andersen,  CB.,  Filopanti,  M.,  Jin,  SL.,  Conti,  M.  (2002)  Cyclic  nucleotidephosphodiesterases  and  their  role  in  endocrine  cell  signaling.  Trends  in  Endocrinology  &Metabolism 13, 29­35.Meisel,  H.  (2001) Bioactive  peptides  from  milk  proteins:  a  perspective  for  consumers  andproducers. Austr. J. Dairy Technol. 56:83­91.Meisel, H.,  FitzGerald,  R.J.  (2003)  Biofunctional  peptides  from  milk  proteins:  mineralbinding and cytomodulatory effects. Curr Pharm Des 9, 1289­1295.Meisel,  H.,  Bernard,  H.,  Fairweather­Tait,  S.,  FitzGerald,  R.J.,  Hartmann,  R.,  Lane,C.N., McDonagh, D., Teucher, B., Wal, J.M. (2003) Detection of caseinophosphopeptidesin the distal ileostomy fluid of human subjects. Br J Nutr 89, 351­359.Mery, P.F., Lohmann, S.M., Walter, U., Fischmeister, R. (1991) Ca2+ current is regulatedby  cyclic GMP­dependent  protein  kinase  in  mammalian  cardiac  myocytes.  Proc  Natl  AcadUSA 88, 1197­201.Michibata,  H.,  Yanaka,  N.,  Kanoh,  Y.,  Okumura,  K.,  Omori,  K.  (2001)  HumanCa2+/calmodulin­dependent  phosphodiesterase  PDE1A:  novel  splice  variants,  their  specificexpression,  genomic  organization,  and  chromosomal  localization.  Biochim  Biophys  Acta1517, 278­87.

81

Michaeli, T., Bloom, T.J., Martins, T., Loughney, K., Ferguson, K., Riggs, M., Rodgers,L., Beavo, J.A., Wigler, M. (1993) Isolation and characterization of a previously undetectedhuman  cAMP phosphodiesterase  by  complementation of  cAMP phosphodiesterase­deficientSaccharomyces cerevisiae. J Biol Chem 268, 12925­32.Miguel, M., Aleixandre, A. (2006) Antihypertensive peptides derived  from egg proteins.  JNutr 136, 1457­1460.Miki, T., Taira, M., Hockman, S., Shimada, F., Lieman, J., Napolitano, M., Ward, D.,Taira, M., Makino, H., Manganiello, V.C. (1996) Characterization of  the cDNA and geneencoding  human  PDE3B,  the  cGIP1  isoform  of  the  human  cyclic  GMP­inhibited  cyclicnucleotide phosphodiesterase family. Genomics 36, 476­85.Miro,  X.,  Perez­Torres,  S.,  Artigas,  F.,  Puigdomenech,  P.,  Palacios,  J.M.,  Mengod,  G.(2002)  Regulation of cAMP phosphodiesterase mRNAs expression in rat brain by acute andchronic  fluoxetine  treatment.  An  in  situ  hybridization  study.  Neuropharmacology  43,  1148­57.Mishaela, R., Rubin, MD., Bilezikian, JP.  (2002) New anabolic  therapies  in osteoporosis.Current Opinion in Rheumatology 2002(14):433­440.Mitsui T., Nomura S., Itakura A., Mizutani S. (2004) Role of aminopeptidases in the bloodpressure regulation. Biol Pharmacoll Bull  27(6):768­71Miyamoto, K., Suzuki, H., Yamamoto, S., Saitoh, Y., Ochiai, E., Moritani, S., Yokogawa,K., Waki, Y., Kasugai, S., Sawanishi, H., Yamagami, H. (2003) Prostaglandin E2­mediatedanabolic  effect  of  a  novel  inhibitor  of  phosphodiesterase  4,  XT­611,  in  the  in  vitro  bonemarrow culture. J Bone Miner Res 18:1471­7.Miyamoto, K., Waki, Y., Horita, T., Kasugai, S., Ohya, K. (1997) Reduction of bone lossby denbufylline, an inhibitor of phosphodiesterase 4. Biochem Pharmacol 54, 613­6177.Miyamoto, K., Suzuki, H., Yamamoto, S., Saitoh, Y., Ochiai, E., Moritani, S., Yokogawa,Ng  KW.  Romas  E.  Donnan  L.  Findlay  DM. (1997) Bone  biology. Baillieres  ClinicalEndocrinology & Metabolism. 11(1):1­22.Mogi,  M.,  Togari,  A.  (2003)  Activation  of  caspases  is  required  for  osteoblasticdifferentiation. J Biol Chem 278, 47477­47482.Mori, K., Shioi, A., Jono, S., Nishizawa, Y., Morii, H. (1999) Dexamethasone enhances Invitro  vascular  calcification  by  promoting  osteoblastic  differentiation  of  vascular  smoothmuscle cells. Arterioscler Thromb Vascul Biol 19, 2112­2118.Mousses, S.,  Caplen,  N.J.,  Cornelison,  R.,  Weaver,  D.,  Basik,  M.,  Hautaniemi,  S.,Elkahloun, A.G., Lotufo, R.A., Choudary, A., Dougherty, E.R., Suh, E., Kallioniemi, O.(2003) RNAi microarray analysis in cultured mammalian cells. Genome Res 13, 2341­2347.Mullally, M.M.,  Meisel,  H.,  FitzGerald,  R.J.  (1996)  Synthetic  peptides  corresponding  toalpha­lactalbumin  and  beta­lactoglobulin  sequences  with  angiotensin­I­converting  enzymeinhibitory activity. Biological Chemistry Hoppe­Seyler 377, 259­26Muradov, K.G.,  Boyd,  K.K.,  Martinez,  S.E.,  Beavo,  J.A.,  Artemyev,  N.O.  (2003) TheGAFa  domains  of  rod  cGMP­phosphodiesterase  6  determine  the  selectivity  of  the  enzymedimerization. J Biol Chem 278, 10594­601.

82

Murray,  E.,  Provvedini,  D.,  Curran,  D.,  Catherwood,  B.,  Sussman,  H., Manolagas,  S.(1987) Characterization of a human osteoblastic osteosarcoma cell  line (SAOS­2) with highbone alkaline phosphatase activity. J Bone Miner Res 2, 231­8.Nakashima,  K.,  Zhou,  X.,  Kunkel,  G.,  Zhang,  Z.,  Deng,  J.M,  Behringer,  R.R.,  deCrombrugghe,  B.  (2002)  The  novel  zing  finger­containing  transcription  factor  osterix  isrequired for osteoblast differentiation and bone formation. Cell 108, 17­29.Nakashima,  A.,  Katagiri,  T.,  Tamura,  M.  (2004)  Cross­talk  between  Wnt  and  bonemorphogenetic protein 2 (BMP­2) signaling in differentiation pathway of C2C12 myoblasts. JBiol Chem 280, 37660­37668.Nakata,  A.,  Ogawa,  K.,  Sasaki,  T.,  Koyama,  N.,  Wada,  K.,  Kotera,  J.,  Kikkawa,  H.,Omori,  K.,  Kaminuma,  O.  (2002)  Potential  role  of  phosphodiesterase  7  in  human  T  cellfunction:  comparative  effects  of  two  phosphodiesterase  inhibitors.  Clin  Exp  Immunol 128,460­466.Narita, M., Murata,  T.,  Shimizu,  K.,  Sugiyama,  T.,  Nakagawa,  T.,  Manganiello,  VC.,Tagawa,  T.  (2003) Phosphodiesterase  4  in  osteoblastic  osteosarcoma  cells  as  a  potentialtarget for growth inhibition. Anti­Cancer Drug. 14, 377­381.Narva, M.,  Halleen,  J.,  Vaananen,  K.,  Korpela,  R.  (2004)  Effects  of  Lactobacillushelveticus fermented milk on bone cells in vitro. Life Sci 75, 1727­1734.Narva,  M.,  Rissanen,  J.,  Halleen,  J.,  Vapaatalo,  H.,  Vaananen, K., Korpela,  R.  (2007)Effects  of  bioactive  peptide,  valyl­prolyl­proline  (VPP),  and  lactobacillus  helveticusfermented milk containing VPP on bone loss in ovariectomized rats. Ann Nutr Metab 51, 65­74.Nelson, W.J., Nusse, R.  (2004) Convergence of Wnt, beta­catenin, and cadherin pathways.Science 303, 1483­1487.Nielsen, C.U.,  Brodin,  B.  (2003)  Di/tri­peptide  transporters  as  drug  delivery  targets:regulation  of  transport  under  physiological  and  patho­physiological  conditions  Curr  DrugTargets 4, 373­388.Noda, M., Yoon, K., Rodan, G.A. (1988) Cyclic AMP­mediated stabilization of osteocalcinmRNA  in  rat  osteoblast­like  cells  treated  with  parathyroid  hormone.  J  Biol  Chem  263,18574–18577.Noda,  M.,  Rodan, G.A.  (1989)  Transcriptional  regulation  of  osteopontin  production  in  ratosteoblast­like cells by parathyroid hormone. J Cell Biol 108, 713­718.Nuttall, M.E.,  Patton,  A.J.,  Olivera,  D.L.,  Nadeau,  D.P.,  Gowen,  M.  (1998)  Humantrabecular  bone  cells  are  able  to  express  both  osteoblastic  and  adipocytic  phenotype:implications for osteopenic disorders. J Bone Miner Res 13, 371­82.O'Brien, C.A., Jia, D., Plotkin, L.I., Bellido, T., Powers, C.C., Stewart, S.A., Manolagas,S.C.,  Weinstein,  R.S.  (2004)  Glucocorticoids  act  directly  on  osteoblasts  and  osteocytes  toinduce  their  apoptosis  and  reduce  bone  formation  and  strength.  Endocrinology  145,  1835­1841.Oger,  S., Mehats,  C.,  Dallot, E.,  Ferre,  F.,  Leroy,  M.J.  (2002) Interleukin­1beta  inducesphosphodiesterase  4B2  expression  in  human  myometrial  cells  through  a  prostaglandin  E2­

83

and cyclic adenosine 3',5'­monophosphate­dependent pathway.  J Clin Endocrinol Metab 87,5524­31.Ondetti, M.A.,  Rubin,  B.,  Cushman,  D.W.  (1977)  Design  of  specific  inhibitors  ofangiotensin­converting  enzyme:  new  class  of  orally  active  antihypertensive  agents.  Science196, 441­444.Oki, N., Takahashi, S.I., Hidaka, H., Conti, M.  (2000) Short  term  feedback  regulation ofcAMP in FRTL­5 thyroid cells. Role of PDE4D3 phosphodiesterase activation. J Biol Chem275, 10831­7.Omi,  K.,  Tokunaga,  K.,  Hohjoh,  H.  (2004)  Long­lasting  RNAi  activity  in  mammalianneurons. FEBS Lett 558, 89­95.Omori, K., Kotera, J. (2007) Overview of PDEs and their regulation. Circ Res 100, 309­327.Pai, S.I., Lin, Y.Y., Macaes, B., Meneshian, A., Hung, C.F., Wu, T.C. (2006) Prospects ofRNA interference therapy for cancer. Gene Ther 13, 464­477.Parfitt,  A.M. (1994) Osteonal  and  hemi­osteonal  remodeling:  the  spatial  and  temporalframework for signal traffic in adult human bone. J Cell Biochem 55(3):273­86.Park,  H., Yim, M.  (2007)  Rolipram,  a  phosphodiesterase  4  inhibitor,  suppresses  PGE2­induced  osteoclast  formation  by  lowering  osteoclast  progenitor  cell  viability.  Arch  PharmaRes 30, 486­492.Partridge, N.C., Bloch, S.R., Pearman, A.T. (1994) Signal transduction pathways mediatingparathyroid hormone regulation of osteoblastic gene expression. J Cell Biochem 55, 321­327.Patschan,  D.,  Loddenkemper,  K.,  Buttgereit,  F.  (2001)  Molecular  mechanisms  ofglucocorticoid­induced osteoporosis. Bone 29(6):498­505.Pavalko,  F.M.,  Otey,  C.A.  (1994) Role  of  adhesion  molecule  cytoplasmic  domains  inmediating interactions with the cytoskeleton. Proc Soc Exp Biol Med 205, 282­293.Pfeifer,  A., Aszodi, A., Seidler,  U., Ruth,  P., Hofmann, F.,  Fassler,  R.  (1996)  Intestinalsecretory defects and dwarfism in mice lacking cGMP­dependent protein kinase II.   Science274, 2082­2086.Pilz, R.B, Broderick K.E.  (2005) Role of cyclic GMP  in gene regulation. Front Biosci 10,1239­1268.Price, P.A., Otsuka, A.A., Poser, J.W., Kristaponis, J., Raman, N. (1976) Characterizationof  a gamma­carboxyglutamic  acid­containing  protein  from  bone. P  Natl  Acad  Sci  USA  73,1447­1451.Rahn, T., Ronnstrand, L., Leroy, M.J., Wernstedt, C., Tornqvist, H., Manganiello, V.C.,Belfrage,  P.,  Degerman,  E.  (1996)  Identification  of  the  site  in  the  cGMP­inhibitedphosphodiesterase  phosphorylated  in  adipocytes  in  response  to  insulin  and  isoproterenol.  JBiol Chem 271, 11575­80.Raisz,  L.G.  (1999)  Physiology  and  pathophysiology  of  bone  remodeling.  Clin  Chem  45,1353­1358.Rawadi,  G.,  Ferrer,  C.,  Spinella­Jaegle,  S.,  Roman­Roman,  S.,  Bouali,  Y.,  Baron,  R.(2001) 1­(5­oxohexyl)­3,7­Dimethylxanthine, a phosphodiesterase inhibitor, activates MAPKcascades  and  promotes  osteoblast  differentiation  by  a  mechanism  independent  of  PKA

84

activation  (pentoxifylline promotes osteoblast differentiation). Endocrinology 142(11):4673­82.Rawadi,  G.,  Roman­Roman,  S.  (2005)  Wnt  signalling pathway:  a  new  target  for  thetreatment of osteoporosis. Expert Opin Ther Tar 9, 1063­1077.Reeve, J. (1996) PTH: a future role in the management of osteoporosis?. J Bone Miner Res11, 440­445.Rehfuss,  R.P.,  Walton,  K.M.,  Loriaux,  M.M.,  Goodman,  R.H.  (1991)  The  cAMP­regulated  enhancer­binding  protein  ATF­1  activates  transcription  in  response  to  cAMP­dependent protein kinase A. J Biol Chem 266, 18431­4.Reinhardt, R.R., Chin, E., Zhou, J., Taira, M., Murata, T., Manganiello, V.C., Bondy,C.A.  (1995) Distinctive  anatomical  patterns  of  gene  expression  for  cGMP­inhibited  cyclicnucleotide phosphodiesterases. J Clin Invest 95, 1528­38.Rentero,  C.,  Monfort,  A.,  Puigdomenech,  P.  (2003)  Identification  and  distribution  ofdifferent  mRNA  variants  produced  by  differential  splicing  in  the  human  phosphodiesterase9A gene. Biochem Biophys Res Commun 301, 686­92.Reyes, M., Verfaillie, C.M. (2001) Characterization of multipotent adult progenitor cells,  asubpopulation of mesenchymal stem cells, Ann. N. Y. Acad. Sci. 938, 231–233.Richter, W., Conti, M. (2002) Dimerization of the type 4 cAMP­specific phosphodiesterasesis mediated by the upstream conserved regions (UCRs). J Biol Chem 277, 40212­21.Rizzoli, R., Bonjour, J.P. (2004) Dietary protein and bone health.[comment]. J Bone MinerRes 19, 527­531.Rizzello, C.G., Losito, I., Gobbetti, M., Carbonara, T., De Bari, M.D., Zambonin, P.G.(2005a) Antibacterial  activities of peptides  from  the water­soluble extracts of  Italian cheesevarieties. J Dairy Sci 88, 2348­2360.Rizzello, C.G., Losito, I., Gobbetti, M., Carbonara, T., De Bari, M.D., Zambonin, P.G.(2005b) Antibacterial activities of peptides  from the water­soluble extracts of Italian cheesevarieties. J Dairy Sci 88, 2348­2360.Rodan,  S.B.,  Imai,  Y.,  Thiede,  M.A.,  Wesolowski,  G.,  Thompson,  D.,  Bar­Shavit,  Z.,Shull, S., Mann, K., Rodan, G.A. (1987) Characterization of a human osteosarcoma cell line(Saos­2) with osteoblastic properties. Cancer Research 47, 4961­6.Rondinone, C.M., Carvalho, E., Rahn, T., Manganiello, V.C., Degerman, E., Smith, U.P.(2000) Phosphorylation  of  PDE3B  by  phosphatidylinositol  3­kinase  associated  with  theinsulin receptor. J Biol Chem 275, 10093­8.Rosen,  C.J.,  Donahue, L.R.  (1998) Insulin  ­like growth  factors  and  bone­the  osteoporosisconnection revisited. Proc Soc Exp Biol Med 219, 1­7.Rosman, G.J., Martins, T.J., Sonnenburg, W.K., Beavo, J.A., Ferguson, K., Loughney,K. (1997) Isolation and characterization of human cDNAs encoding a cGMP­stimulated 3',5'­cyclic nucleotide phosphodiesterase. Gene 191, 89­95.Ross,  E.M.,  Gilman,  A.G.  (1977)  Resolution  of  some  components  of  adenylate  cyclasenecessary for catalytic activity. J Biol Chem 252, 6966­9.Rotella, DP. (2002) Phosphodiesterase 5 inhibitors: current status and potential applications.Nature Reviews. Drug Discovery 1(9):674­82.

85

Sakamoto,  M.K.,  Suzuki,  K.,  Takiya,  S.,  Yoshimura,  Y.,  Imai,  T.,  Matsumoto,  A.,Nakamura,  S.  (1998)  Developmental  profiles  of  phosphorylated  and  unphosphorylatedCREBs in murine calvarial MC3T3­E1 cells. J Biochem 3, 399­407.Sasaki, N., Kusano, E., Ando, Y., Yano, K., Tsuda, E., Asano, Y. (2001) Glucocorticoiddecreases circulating osteoprotegerin (OPG): possible mechanism for glucocorticoid inducedosteoporosis. Nephrol Dial Transplant 16, 479­482.Sasaki,  T., Kotera,  J.,  Yuasa,  K.,  Omori,  K.  (2000)  Identification  of  human  PDE7B,  acAMP­specific phosphodiesterase. Biochem Bioph Res Co 271, 575­83.Sassone­Corsi, P. (1995) Transcription factors responsive to cAMP. Annu Rev Cell Dev Biol11, 355­77.Schaffner,  P.,  Meyer,  J.,  Dard,  M.,  Wenz,  R.,  Nies,  B.,  Verrier,  S.,  Kessler,  H.,Kantlehner,  M.  (1999)  Induced  tissue  integration  of  bone  implants  by  coating  with  boneselective RGD­peptides in vitro and in vivo studies. J Mater Scie Mater Med 10, 837­839.Scherr, M., Eder, M. (2007) Gene silencing by small regulatory RNAs in mammalian cells.Cell Cycle 6, 444­9.Schuler, M.,  Owen,  G.R.,  Hamilton,  D.W.,  de  Wild,  M.,  Textor,  M.,  Brunette,  D.M.,Tosatti,  S.G.  (2006)  Biomimetic  modification  of  titanium  dental  implant  model  surfacesusing the RGDSP­peptide sequence: a cell morphology study. Biomaterials 27, 4003­4015.Scutt,  A.,  Beier,  N.,  Fittschen,  C.  (2004)  EMD273316  &  EMD95833,  type  4phosphodiesterase  inhibitors,  stimulate  fibroblastic­colony  formation  by  bone  marrow  cellsvia direct inhibition of PDE4 and the induction of endogenous prostaglandin synthesis. BMCPharmacology 4(1);10.Semizarov, D.,  Kroeger,  P.,  Fesik,  S.  (2004)  siRNA­mediated  gene  silencing:  a  globalgenome view. Nucleic Acids Res 32, 3836­3845.Sette,  C.,  Iona,  S.,  Conti,  M.  (1994) The  short­term  activation  of  a  rolipram­sensitive,cAMP­specific phosphodiesterase by thyroid­stimulating hormone in thyroid FRTL­5 cells ismediated by a cAMP­dependent phosphorylation. J Biol Chem 269, 9245­52.Sette,  C.,  Conti,  M.  (1996)  Phosphorylation  and  activation  of  a  cAMP­specificphosphodiesterase  by  the  cAMP­dependent  protein  kinase.  Involvement  of  serine  54  in  theenzyme activation. J Biol Chem 271, 16526­34.Shalhoub, V., Conlon, D., Tassinari, M., Quinn, C., Partridge, N., Stein, G.S., Lian, J.B.(1992) Glucocorticoids  promote  development  of  the  osteoblast  phenotype  by  selectivelymodulating expression of cell growth and differentiation associated genes. J Cell Biochem 50,425­440.Shi,  X.M.,  Blair,  H.C.,  Yang,  X.,  McDonald,  J.M.,  Cao,  X.  (2000)  Tandem  repeat  ofC/EBP  binding  sites  mediates  PPARgamma2  gene  transcription  in  glucocorticoid­inducedadipocyte differentiation. J Cell Biochem 76, 518­527.Shimabukuro, M., Zhou, Y­T., Levi, M., Unger, RH.  (1998) Fatty­acid  induced beta cellapoptosis: a link between obesity and diabetes. Proc Natl Acad Sci USA 95, 2498­2501.Simonet, W.S., Lacey, D.L., Dunstan, C.R., Kelley, M., Chang, M.S., Luthy, R., Nguyen,H.Q.,  Wooden,  S.,  Bennett,  L.,  Boone,  T.,  Shimamoto,  G.,  DeRose,  M.,  Elliott,  R.,Colombero, A., Tan, H.L., Trail, G., Sullivan, J., Davy, E., Bucay, N., Renshaw­Gegg, L.,

86

Hughes,  T.M.,  Hill,  D.,  Pattison,  W.,  Campbell,  P.,  Sander,  S.,  Van,  G.,  Tarpley,  J.,Derby, P., Lee, R., Boyle, W.J. (1997) Osteoprotegerin: a novel secreted protein involved inthe regulation of bone density. Cell 89, 309­319.Skalhegg,  B.S.,  Tasken,  K.  (2000)  Specificity  in  the  cAMP/PKA  signaling  pathway.Differential  expression,regulation,  and  subcellular  localization  of  subunits  of  PKA.  FrontBiosci 5, D678­693.Smith,  S.J.,  Brookes­Fazakerley,  S.,  Donnelly,  L.E.,  Barnes,  P.J.,  Barnette,  M.S.,Giembycz,  M.A.  (2003)  Ubiquitous  expression  of  phosphodiesterase  7A  in  humanproinflammatory and immune cells. Am J Physiol Lung Cell Mol Physiol 284, L279­89.Soderling, S.H., Bayuga, S.J., Beavo, J.A. (1998a) Cloning and characterization of a cAMP­specific cyclic nucleotide phoshodiesterase. Proc Natl Acad Sci USA 95, 8991­8996.Soderling, S.H., Bayuga, S.J., Beavo, J.A. (1998b)  Identification and characterization of anovel family of cyclic nucleotide phoshodiesterases. J Biol Chem 273, 15553­15558.Soderling,  S.H., Beavo,  J.A.  (2000) Regulation  of  cAMP  and  cGMP  signaling:  newphosphodiesterases and new functions. Curr Opin Cell Biol 12,174­9.Sonnenburg, W.K., Mullaney, P.J., Beavo, J.A. (1991) Molecular cloning of a cyclic GMP­stimulated  cyclic  nucleotide  phosphodiesterase  cDNA.  Identification  and  distribution  ofisozyme variants. J Biol Chem 266, 17655­61.Sooy,  K.,  Sabbagh,  Y.,  Demay,  M.B.  (2005)  Osteoblasts  lacking  the  vitamin  D  receptordisplay enhanced osteogenic potential in vitro. J Cell Biochem 94, 81­87.Sprague,  S.M.,  Popovtzer,  M.M.,  Dranitzki­Elhalel,  M.,  Wald,  H.  (1996)  Parathyroidhormone­induced  calcium  efflux  from  cultured  bone  is  mediated  by  protein  kinase  Ctranslocation. Am J Physiol 271, F1139­46.Spence, S., Rena, G., Sullivan, M., Erdogan, S., Houslay, M.D. (1997) Receptor­mediatedstimulation of  lipid signalling pathways  in CHO cells elicits the rapid transient  induction ofthe  PDE1B  isoform  of  Ca2+/calmodulin­stimulated  cAMP  phosphodiesterase.  Biochem  J321, 157­63.Stanford, C.M., Jacobson, P.A., Eanes, E.D., Lembke, L.A., Midura, R.J. (1995) RapidlyForming Apatitic Mineral in an Osteoblastic Cell Line (UMR 106­01 BSP). J Biol Chem 270,9420–9428.Steffansen,  B., Nielsen, C.U.,  Brodin,  B.,  Eriksson,  A.H.,  Andersen,  R.,  Frokjaer,  S.(2004) Intestinal solute carriers: an overview of trends and strategies for improving oral drugabsorption. Eur J Pharm Sci 21, 3­16.Stork, P.J., Schmitt, J.M. (2002) Crosstalk between cAMP and MAP kinase signaling in theregulation of cell proliferation. Trends Cell Biol 12,  258­266.Swarthout, J.T.,  Doggett,  T.A.,  Lemker,  J.L.,  Partridge,  N.C.  (2001)  Stimulation  ofextracellular signal­regulated kinases and proliferation in rat osteoblastic cells by parathyroidhormone is protein kinase C­dependent. J Biol Chem 276, 7586­7592.Swarthout, J.T.,  D'Alonzo,  R.C.,  Selvamurugan,  N.,  Partridge,  N.C.  (2002)  Parathyroidhormone­dependent signaling pathways regulating genes in bone cells. Gene 282, 1­17.

87

Swinnen, J.V., Tsikalas, K.E., Conti, M. (1991) Properties and hormonal regulation of twostructurally  related  cAMP  phosphodiesterases  from  the  rat  Sertoli  cell.  J  Biol  Chem  266,18370­7.Takabatake, Y., Isaka, Y., Mizui, M., Kawachi, H., Shimizu, F., Ito T., Hori, M., Imai, E.(2005) Exploring RNA interference as a therapeutic strategy for renal disease. Gene Ther 12,965­973.Takami, M., Cho, E.S., Lee, S.Y., Kamijo, R., Yim, M. (2005) Phosphodiesterase inhibitorsstimulate  osteoclast  formation  via  TRANCE/RANKL  expression  in  osteoblasts:  possibleinvolvement of ERK and p38 MAPK pathways. FEBS Lett 579,  832­838.Takeda, K., Kaisho, Y., Yoshida, N., Takeda, J., Kishimoto, T., Akira, S. (1998) STAT3activation is responsible for IL­6­dependent T cell proliferation through preventing apoptosis:generation  and  characterizaion  of  T  cell­specific  STAT3­deficient  mice.  J  Immunol  161,4652­4660.Tam,  C.S.,  Heersche,  J.N.,  Murray,  T.M.,  Parsons,  J.A.  (1982)  Parathyroid  hormonestimulates  the bone apposition rate independently of  its resorptive action: differential effectsof intermittent and continuous administration. Endocrinology 110, 506­512.Tanaka, H., Seino, Y. (2004) Direct action of 1,25­dihydroxyvitamin D on bone: VDRKObone shows excessive bone formation in normal mineral condition. J Steroid Biochem 89­90,343­345.Taussig,  R.,  Gilman,  A.G.  (1995)  Mammalian  membrane­bound  adenylyl  cyclases,  J  BiolChem 270, 1–4.Taylor,  S.S.,  Kim,  C.,  Vigil,  D.,  Haste,  N.M.,  Yang,  J.,  Wu,  J.,  Anand,  G.S.  (2005)Dynamics of signaling by PKA. Biochimi Biophys Acta 1754, 25­37.Teitelbaum, S.L. (2000) Bone resorption by osteoclast. Science 289, 1504­1508.Teitelbaum SL. Ross FP. (2003) Genetic regulation of osteoclast development and function.Nature Reviews Genetics  4(8), 638­49.Theofan, G., Price, P.A. (1989) Bone Gla protein messenger ribonucleic acid is regulated byboth  1,25­dihydroxyvitamin  D3  and  3',5'­cyclic  adenosine  monophosphate  in  ratosteosarcoma cells. Mol Endocrinol 3, 36­43.Thompson,  W.J.,  Appleman,  M.M.  (1971)  Multiple  cyclic  nucleotide  phosphodiesteraseactivities from rat brain. Biochemistry 10, 311­316.Toranzo, G.S., Oterino, J., Zelarayan, L., Bonilla, F., Buhler, M.I. (2007) Spontaneous andLH­induced maturation  in Bufo arenarum oocytes:  importance of gap  junctions.  Zygote 15,65­80.Torphy, T.J., Zhou, H.L., Foley, J.J., Sarau, H.M., Manning, C.D., Barnette, M.S. (1995)Salbutamol  up­regulates PDE4  activity  and  induces  a  heterologous desensitization  of  U937cells to prostaglandin E2. Implications for the therapeutic use of beta­adrenoceptor agonists. JBiol Chem 270, 23598­604.Torphy, TJ. (1998)  Phosphodiesterase  isozymes:  molecular  targets  for  novel  antiasthmaagents.  American Journal of Respiratory & Critical Care Medicine 157(2):351­70.Torras­Llort,  M.,  Azorin,  F.  (2003)  Functional  characterization  of  the  humanphosphodiesterase 7A1 promoter. Biochem J 373, 835­43.

88

Tsuboi, Y., Shankland, S.J., Grande, J.P., Walker, H.J., Johnson, R.J., Dousa, T.P. (1996)Suppression of  mesangial proliferative glomerulonephritis development  in  rats by  inhibitorsof cAMP phosphodiesterase isozymes types III and IV. J Clin Invest 98, 262­270.Tsou,  K.,  Snyder,  G.L.,  Greengard,  P.  (1993)  Nitric  oxide/cGMP  pathway  stimulatesphosphorylation  of  DARPP­32,  a  dopamine­  and  cAMP­regulated  phosphoprotein,  in  thesubstantia nigra. Proc Natl Acad Sci USA 90, 3462­5.Tuder, R.M., Flook, B.E., Voelkel, N.F. (1995) Increased gene expression for VEGF and theVEGF  receptors  KDR/Flk  and  Flt  in  lungs  exposed  to  acute  or  to  chronic  hypoxia.Modulation of gene expression by nitric oxide. J Clin Invest. 95, 1798­807.Turowski,  P.,  Adamson,  P., Greenwood,  J.  (2005) Pharmacological  targeting  of  ICAM­1signaling  in  brain  endothelial  cells:  potential  for  treating  neuroinflammation.  Cell  MolNeurobiol 25, 153­170.Urena, P.,  Iida­Klein,  A.,  Kong,  X.F.,  Juppner,  H.,  Kronenberg,  H.M.,  Abou­Samra,A.B.,  Segre,  G.V.  (1994)  Regulation  of  parathyroid  hormone  (PTH)/PTH­related  peptidereceptor messenger ribonucleic acid by glucocorticoids and PTH in ROS 17/2.8 and OK cells.Endocrinology 134, 451­456.Uzel, M.I., Scott, I.C., Babakhanlou­Chase, H., Palamakumbura, A.H., Pappano, W.N.,Hong,  H­H.,  Greenspan,  D.S.,  Trackman,  P.C.  (2001)  Multiple  bone  morphogeneticprotein  1­related  mammalian  metalloproteinases  process  pro­lysyl  oxidase  at  the  correctphysiological site and control lysyl oxidase activation in mouse embryo fibroblast cultures. JBiol Chem 276, 22537­22543.Van Sande, J., Lefort, A., Beebe, S., Roger, P., Perret, J., Corbin, J., Dumont, J.E. (1989)Pairs of cyclic  AMP analogs,  that are  specifically  synergistic  for  type  I and  type  II cAMP­dependent  protein  kinases,  mimic  thyrotropin  effects  on  the  function,  differentiationexpression and mitogenesis of dog thyroid cells. Eur J Biochem 183,699­708.van Staveren,  W.C.,  Glick,  J.,  Markerink­van  Ittersum,  M.,  Shimizu, M., Beavo,  J.A.,Steinbusch,  H.W.,  de  Vente,  J.  (2002)  Cloning  and  localization  of  the  cGMP­specificphosphodiesterase type 9 in the rat brain. J Neurocytol 31, 729­41.Veeman,  M.T.,  Axelrod,  J.D.,  Moon,  R.T.  (2003)  A  second  canon.  Functions  andmechanisms of beta­catenin­independent Wnt signaling. Dev Cell 5, 367­377.Verma, U.N.,  Surabhi  RM.  Schmaltieg,  A.,  Becerra,  C.,  Gaynor,  R.B.  (2003) Smallinterfering RNAs directed against beta­catenin inhibit the in vitro and in vivo growth of coloncancer cells. Clin Cancer Res 9,  1291­1300.Verrier, S., Pallu, S., Bareille, R., Jonczyk, A., Meyer, J., Dard, M., Amedee, J.  (2002)Function  of  linear  and  cyclic  RGD­containing  peptides  in  osteoprogenitor  cells  adhesionprocess. Biomaterials 23, 585­596.Zhang,  M.,  Xuan,  S.,  Bouxsein,  M.L.,  von  Stechow,  D.,  Akeno,  N.,  Faugere,  M.C.,Malluche,  H., Zhao, G., Rosen,  C.J.,  Efstratiadis,  A., Clemens,  T.L.  (2002)  Osteoblast­specific  knockout of  the  insulin­like growth  factor  (IGF)  receptor  gene  reveals  an  essentialrole of IGF signaling in bone matrix mineralization. J Biol Chem 277, 44005­12.Zhao,  A.Z.,  Bornfeldt,  K.E.,  Beavo,  J.A.  (1998)  Leptin  inhibits  insulin  secretion  byactivation of phosphodiesterase 3B. J Clin Invest 102, 869­73.

89

Zhao, A.Z., Huan, J.N., Gupta, S., Pal, R., Sahu, A. (2002) A phosphatidylinositol 3­kinasephosphodiesterase 3B­cyclic AMP pathway in hypothalamic action of  leptin on feeding. NatNeurosci 5, 727­8.Zhao,G.,  Monier­Faugere,  M.C.,  Langub,  M.C.,  Geng,  Z.,  Nakayama,  T.,  Pike,  J.W.,Chernausek,  S.D.,  Rosen,  C.J.,  Donahue,  L.R.,  Malluche,  H.H.,  Fagin,  J.A.,  Clemens,T.L.  (2000)  Targeted  overexpression  of  insulin­like  growth  factor  I  to  osteoblasts  oftransgenic mice: increased trabecular bone volume without increased osteoblast proliferation.Endocrinology 141, 2674­82.Zhao, S., Zhang, Y.K., Harris, S., Ahuja, S.S., Bonewald, L.F. (2002) MLO­Y4 osteocyte­like cells support osteoclast formation and activation.  J Bone Miner Res 17, 2068­2079.Zhong,  Y.  (1995)  Mediation  of  PACAP­like  neuropeptide  transmission  by  coactivation  ofRas/Raf and cAMP signal transduction pathways in Drosophila. Nature 375, 588­92.Ziff,  EB.  (1990)  Transcription  factors:  a  new  family  gathers  at  the  cAMP  response  site.Trends Genet 6, 69­72.Zmuda­Trzebiatowska, E., Manganiello, V., Degerman, E.  (2007) Novel  maechanism ofthe  regulation  of  Protein  kinase  B  in  adipocytes;  implications  for  Protein  kinase  A,  Epac,Phoshpodiesterases 3 and 4. Cell Signal 19, 81­86.Zoraghi,  R.,  Bessay,  E.P.,  Corbin,  J.D.,  Francis,  S.H.  (2005) Structural  and  functionalfeatures  in  human  PDE5A1  regulatory  domain  that  provide  for  allosteric  cGMP  binding,dimerization, and regulation. J Biol Chem 280, 12051­63.Yamagami, H.,  Nishioka,  T.,  Ochiai,  E.,  Fukushima,  K.,  Nomura,  M.,  Kasugai,  S.,Moritani,  S.,  Yokogawa,  K.,  Miyamoto,  K.  (2003)  Inhibition  of  osteoclastogenesis  by  aphosphodiesterase  4  inhibitor  XT­611  through  synergistic  action  with  endogenousprostaglandin E2. Biochem Pharmacol 66, 801­807.Yan,  C.,  Zhao,  A.Z.,  Sonnenburg,  W.K.,  Beavo,  J.A.  (2001)  Stage  and  cell­specificexpression  of  calmodulin­dependent  phosphodiesterases  in  mouse  testis.  Biol  Reprod  64,1746­54.Yanaka,  N.,  Kurosawa,  Y.,  Minami,  K.,  Kawai,  E.,  Omori,  K.  (2003)  cGMP­phosphodiesterase activity  is up­regulated  in response to pressure overload of rat ventricles.Biosci Biochem 67, 973­9.Yang, G., McIntyre, K.W., Townsend, R.M., Shen, H.H., Pitts, W.J., Dodd, J.H., Nadler,S.G.,  McKinnon,  M.,  Watson,  A.J.  (2003)  Phosphodiesterase  7A­deficient  mice  havefunctional T cells. J Immunol 171,  6414­6420.Yang,  Q.,  Paskind,  M.,  Bolger,  G.,  Thompson,  W.J.,  Repaske,  D.R.,  Cutler,  L.S.,Epstein,  P.M.  (1994)  A  novel  cyclic  GMP  stimulated  phosphodiesterase  from  rat  brain.Biochem Biophys Res Commun 205, 1850­8.Yuan, J.,  Kramer,  A.,  Matthess,  Y.,  Yan,  R.,  Spankuch,  B.,  Gatje,  R.,  Knecht,  R.,Kaufmann,  M.,  Strebhardt,  K.  (2006)  Stable  gene  silencing  of  cyclin  B1  in  tumor  cellsincreases susceptibility to taxol and leads to growth arrest in vivo. Oncogene 25, 1753­1762.Xing,  L.,  Boyce,  BF.  (2005)  Regulation  of  apoptosis  in  osteoclasts  and  osteoblastic  cells.Biochem Biophys Res Commun 328, 709­720.

90

Wakabayashi, S., Tsutsumimoto, T., Kawasaki, S., Kinoshita, T., Horiuchi, H., Taoka,K. (2002) Involvement of Phoshodiesterase Isozymes  in Osteoblastic Differentiation. J BoneMiner Res 12, 172­178.Waki,  Y.,  Horita,  T.,  Miyamoto,  K.,  Ohya,  K.,  Kasugai,  S.  (1999)  Effects  of  XT­44,  aphosphodiesterase  4  inhibitor,  in  osteoblastgenesis  and  osteoclastgenesis  in  culture  and  itstherapeutic effects Japanese Journal of Pharmacology. 79(4):477­83.Walsh,  C.A.,  Bowler,  W.B.,  Bilbe,  G.,  Fraser,  W.D.,  Gallagher,  J.A.  (1997)  Effects  ofPTH on PTHrP gene expression  in human osteoblasts: up­regulation with the kinetics of animmediate early gene. Biochem Biophys Res Commun 239, 155­159.Valta, M.P., Hentunen, T., Qu, Q., Valve, E.M., Harjula, A., Seppanen, J.A., Vaananen,H.K.,  Harkonen,  P.L.  (2006)  Regulation  of  osteoblast  differentiation:  a  novel  function  forfibroblast growth factor 8. Endocrinology. 147, 2171­2182.Wang, D., Deng, C., Bugaj­Gaweda, B., Kwan, M., Gunwaldsen, C., Leonard, C., Xin,X.,  Hu,  Y.,  Unterbeck,  A.,  De  Vivo,  M.  (2003)  Cloning  and  characterization  of  novelPDE4D isoforms PDE4D6 and PDE4D7. Cell Signal 15, 883­91.Wang,  J.H.,  Sharma,  R.K.,  Mooibroek,  M.J.  (1990) Calmodulin­stimulated  cyclicnucleotide  phosphodiesterases.  In  J.  Beavo,  &  M.D.  Houslay  (Eds.),  Cyclic  NucleotidePhosphodiesterases: Structure, Regulation and Drug Action (pp.19­59). John Wiley and SonsLtd.Wang, P., Wu, P., Ohleth, K.M., Egan, R.W., Billah, M.M. (1999) Phosphodiesterase 4B2is  the  predominant  phosphodiesterase  species  and  undergoes differential  regulation  of  geneexpression in human monocytes and neutrophils. Mol Pharmacol 56, 170­4.Wang, P., Wu, P., Egan, R.W., Billah, M.M. (2003) Identification and characterization of anew  human  type  9  cGMP­specific phosphodiesterase  splice  variant  (PDE9A5).  Differentialtissue distribution and subcellular localization of PDE9A variants. Gene 314, 15­27.Wechsler,  J.,  Choi,  Y.H.,  Krall,  J.,  Ahmad,  F.,  Manganiello,  V.C.,  Movsesian,  M.A.(2002)  Isoforms of cyclic  nucleotide phosphodiesterase PDE3A  in cardiac  myocytes. J BiolChem 277, 38072­8.Weinstein,  R.S.,  Jilka,  R.L.,  Parfitt,  A.M.,  Manolagas,  S.C.  (1998)  Inhibition  ofosteoblastogenesis  and  promotion  of  apoptosis  of  osteoblasts  and  osteocytes  byglucocorticoids. Potential mechanisms of their deleterious effects on bone. J Clin Invest  102,274­82.White,  R.E.,  Kryman,  J.P.,  El­Mowafy,  A.M.,  Han,  G.,  Carrier,  G.O.  (2000)  cAMP­dependent  vasodilators  cross­activate  the  cGMP­dependent  protein  kinase  to  stimulateBK(Ca) channel activity in coronary artery smooth muscle cells. Circ Res 86, 897­905.Whitfield,  J.F., Morley, P. (1995) Small bone­building fragments of parathyroid hormone:new therapeutic agents for osteoporosis. Trends Pharmacol Sci 16, 382­386.Whyte, M.P.,  Obrecht, S.E.,  Finnegan,  P.M., Jones,  J.L.,  Podgornik,  M.N.,  McAlister,W.H.,  Mumm,  S.  (2002)  Osteoprotegerin  deficiency  and  juvenile  Paget's  disease.    NewEngland Journal of Medicine 347, 175­84.Wozney, J.M. (2002) Overview of Bone Morphogenetic Proteins. Spine 27, 52­58.

Erratum sheet

Page 17- line 34, Hentunen et al. 1995 not 1998Page 20- line 28, GCs not GSsPage 57- in Figure 8 (A) and publication II (manuscript), page 25- in Figure 6 (A)

(A)

Control

siRNAPDE7

H-89

siRNAPDE7 +

H-89

0

100

200

% C

hang

e

**

*** ***

Should be

(A)

hMSC

Contro

l

siRNAPDE7

siRNAPDE8A

0

250

500

750

1000

**

bALP

ng/

mg

prot

ein

Page 73 – Reference Conti and Beavo (2007) should be on page 69Page 90 – Reference Valta et al. (2006) should be on page 88