Tcpik(bim�c©

Nrmkllpcocct

ddesrbdoscobUrrmtmi

S

Kidney-Bone, Bone-Kidney, and Cell-Cell Communicationsin Renal Osteodystrophy

By Keith A. Hruska, Georges Saab, Lala R. Chaudhary, Cheryl O. Quinn, Richard J. Lund,and Kameswaran Surendran

he relationship between bone and the kidney in renal osteodystrophy is a complex interplay of kidney to boneonnections, bone to kidney connections, and cell to cell connections. In addition, such interactions have arofound effect on the vasculature. In this review, we discuss the role of the bone morphogenetic proteins (BMPs)

n the skeleton, kidney, and vasculature. In addition, we propose that deficiencies of these BMPs seen in chronicidney disease (CKD) result in decreased bone remodeling and a compensatory secondary hyperparathyroidism

high turnover state). Treatment of the hyperparathyroidism blocks this compensatory arm and thus decreasedone remodeling occurs (low turnover). We review animal models of CKD in which treatment with BMP-7 resulted

n normalization of both high and low turnover states. Finally, we discuss vascular calcification as it relates to boneetabolism. We discuss the roles of BMP-7 and 2 other bone regulatory proteins, osteoprotegerin (OPG) and2-HS glycoprotein (AHSG, human fetuin), in the human vasculature and their implications for vascularalcification.

2004 Elsevier Inc. All rights reserved.

aumbidsbtvotkrsmdt

dap

Rs

m5L

EW DISCOVERIES RELATED to sub-stances made in the kidney regulating bone

emodeling are addressed in this article along withention of new skeletal hormones that regulate

idney function. In addition, recent progress re-ated to the receptor activator of nuclear factor � Bigand (RANKL)/RANK/osteoprotegerin (OPG)athway, the bone regulatory protein �2-HS gly-oprotein, and the dependency of bone remodelingn heterotypic cell-cell communications is dis-ussed. Thus, kidney-bone, bone-kidney, and cell-ell communications are discussed as they relate tohe pathogenesis of renal osteodystrophy.

BONE REMODELING

The human skeleton is remodeled continuouslyuring life, and the rate of remodeling is stimulateduring growth and fracture healing. A conceptxists that skeletal growth is a modeling-only is-ue. This is not true because the term modelingefers to the function of the epiphyses of longones where their elongation occurs owing to en-ochondral bone formation and the process of peri-steal deposition and endosteal resorption thathape long bones. However, a critical remodelingomponent is present during growth in most partsf the skeleton. At remodeling sites, the rates ofone formation exceed those of bone resorption.nappreciated until the present time are critical

egulatory relationships between bone modeling,emodeling, and the hematopoietic system in thearrow space of long bones. These ties contribute

o the normal function of the alternate environ-ent. Thus, the functioning immune system is an

mportant regulator of bone remodeling (Fig 1),

eminars in Nephrology, Vol 24, No 1 (January), 2004: pp 25-38

nd the bone mesenchymal compartment contrib-tes to regulation of hematopoiesis. Bone is theajor storehouse of calcium, phosphorus, and

uffer equivalents in the body, and bone remodel-ng is regulated to maintain calcium homeostasisuring pregnancy, lactation, and metabolic acido-is. Osteodystrophies develop from disorderedone remodeling—such is the case in renal os-eodystrophy.1-4 Regulation of bone remodeling in-olves both endocrine and paracrine signaling tosteoblasts and osteoclasts through hormones, cy-okines, and growth factors (Fig 2). In chronicidney disease (CKD), aberrant levels of agentsegulating bone remodeling are released into theystem, resulting in the lack of normal bone for-ation rates, inappropriate bone resorption, and

efects in mineralization constituting renal os-eodystrophy.

THE KIDNEY-BONE AND BONE-KIDNEYCONNECTIONS

Evidence exists to indicate that CKD produces aecrease in skeletal remodeling that leads to anttempt to maintain remodeling through increasedarathyroid hormone (PTH) activity. The best clin-

From the Renal Division, Departments of Pediatrics, and theenal Division, Department of Medicine, Washington Univer-ity School of Medicine, St. Louis, MO.Address reprints requests to Keith A. Hruska, MD, Depart-ent of Pediatrics, Washington University School of Medicine,th Floor MPRB, Campus Box 8208, 660 S. Euclid Ave, St.ouis, Mo 63110. Email: hruska_k@wustl.edu© 2004 Elsevier Inc. All rights reserved.0270-9295/04/2401-0005$30.00/0

doi:10.1053/j.semnephrol.2003.08.010

25

iptrtaccadtcifvmio

acb

T

fidABticaf

srp6foctbRrib

HRUSKA ET AL26

cal demonstration of this is the observation thatrevention of hyperparathyroidism in CKD pa-ients leads to the adynamic bone disorder.3 Weecently have shown in an animal model of CKDhat maintenance of normal phosphorous, calcium,nd PTH levels through phosphate restriction andalcitriol supplementation leads to a marked de-rease in osteoblast number, bone formation rates,nd activation frequency similar to adynamic boneisease.4 Based on these findings, we hypothesizedhat CKD inhibits osteoblast differentiation by de-reased growth factors and/or increased levels ofnhibitory substances blocking the action of growthactors. Two important regulatory families in-olved in osteoblast differentiation are the boneorphogenetic proteins (BMPs) and the Wnt fam-

ly proteins. Phosphatonins are potential inhibitors

Fig 1. Interactions between systems and cell typestimulated by hormones such as PTH (interaction beemodeling), interleukin 1 and TNF�, cytokines releaseroduced by cells of the immune system contribute ma, which support osteoclast progenitor survival and diormation and bone resorption because differentiatisteoblast lineage, and preosteoblasts directly inteell–expressed RANKL binding to its receptor, RANK,erminal differentiation and regulation of osteoclast fuy bone cells and immune cells, decreases RANKL stimANK. The effects of PTH and calcitriol (1,25[OH]2

esorption are mediated by their stimulation of RANKmmune cell activation leading to increased interleukione marrow environment.

f the Wnt family and are discussed in a separate o

rticle in this issue by Kumar. We focus our dis-ussion on the BMPs and their relationship toone, the kidney, and the vasculature.

he BMPs in Bone

BMPs are members of the transforming growthactor � (TGF-�) superfamily. BMPs were firstsolated from demineralized bone, indicating thatemineralized bone is a rich source for BMPs.5,6

t least 15 BMPs have been identified, of whichMP-2, BMP-3, BMP-4, BMP-6, and BMP-7 (os-

eogenic protein-1) have been shown to be potentnducers of ontogenesis. BMPs interact with aomplex set of type I and type II receptors of thectivin receptor family to mediate their biologicunction7,8 through activation of the Smad family

stimulation of bone remodeling. Remodeling can bethe system for divalent ion homeostasis and bonecells of the immune system. Furthermore, cytokines

age colony stimulating factor (M-CSF) and interleukiniation. Bone remodeling is a coupled process of bonebone marrow stromal cells, the progenitors of theith cells in the osteoclast lineage through surfacepreosteoclasts. RANKL is required and sufficient forOsteoprotegerin, a RANKL decoy receptor producedn by decreasing availability of RANKL for binding withosteoclast differentiation and stimulation of bone

duction by preosteoblasts. CKD is a state of alteredF�, and interleukin 6 levels in the circulation and the

duringtweend by Tcroph

fferenton ofract won the

nction.ulatio

D3) onL pro

n 1, TN

f transcription factors.

gcuctg

aopgnaoBBeBeiei

im

daTrBtmfPIppsmacerc

srhleaa ells is

RENAL OSTEODYSTROPHY 27

BMPs have a profound effect on osteoblasticrowth and differentiation as shown by in vitro cellulture studies. BMP-2, BMP-4, and BMP-6 stim-late osteoblast differentiation of rat and mouseells.9-11 BMP-2 and BMP-7 stimulate a differen-iation program of human bone marrow osteopro-enitors and bone-derived osteoblasts.12-15

This is accompanied by increased expression oflkaline phosphatase, type I collagen, osteopontin,steocalcin, decorin, osteonectin, and bone sialo-rotein. BMP-2 is a potent inducer of osteopro-enitor differentiation with a decreased effective-ess in more mature cells. BMP-3 increasedlkaline phosphatase activity, type I collagen, andsteocalcin in human bone marrow stromal cells.16

MP-3 antagonizes the osteogenic effects ofMP-2 and TGF-� in osteoblastic cells.17,18 Ad-noviral vectors carrying BMP-2, BMP-4, BMP-6,MP-7, and BMP-13 complementary DNAs deliv-red intramuscularly or subcutaneously to animalsnduce ectopic bone formation at the site of deliv-ry through endochondral ossification,19-21 show-

Fig 2. The physiology and pathophysiology of osteotem cells (also referred to as bone marrow stromaegulated by the BMP family and the Wnt family of diffeormonal origin deriving from the kidney. PTH regulat

evels by stimulating osteoprogenitor proliferation andffects of PTH on the phenotype of cells in the osteobffects osteoblast differentiation by impairing the BMPre increased, but the phenotype of the osteoblastic c

ng that BMP gene therapy may have potential use a

n the treatment of degenerative, rheumatic, trau-atic bone injury, and kidney diseases.The expression of BMP-7 in adult rat kidney is

ecreased in conditions such as animal models ofcute renal ischemia and diabetic nephropathy.22-25

his suggests that part of the pathophysiology ofenal osteodystrophy may in part be caused byMP deficiency and decreased osteoblast differen-

iation (Fig 2). Hyperparathyroidism, an adaptiveechanism to maintain remodeling rates in the

ace of CKD, is a maladaptive process becauseTH is not an osteoblast differentiation factor.26-30

ncreased levels of PTH stimulate an abnormalhenotype of osteoblastic cells with fibroblast-likeroperties that accumulate in the peritrabecularpace and produce marrow fibrosis (Fig 3). Treat-ent with recombinant BMP-7 results in the dis-

ppearance of these cells, probably by promotingommitment to the osteoblast program.31 BMP-7,ven in the presence of high PTH levels, stimulatedemodeling, increased bone formation, and de-reased bone resorption. BMP-7 also reversed the

differentiation. Osteoblasts derive from mesenchymal; Fig 1) through a multistep differentiation programtion factors. Some of the skeletal BMP-7 burden is ofoblast differentiation in the presence of normal BMPing mature osteoblast and osteocytes apoptosis. The

ifferentiation program are resorptive (see Fig 3). CKDnt physiologic system. As an adaptation, PTH levels

not normal and excess bone resorption is stimulated.

blastl cellsrentia

es osteinhibitlast dand W

dynamic bone disorder and restored bone remod-

etmcBt

T

aIoitmtdh

ardattstkt

iepvBimba

B

tmac(gfiVBsevmvla

HRUSKA ET AL28

ling.4 The successful treatment of renal osteodys-rophy (both high and low turnover) in animalodels of CKD, regardless of serum PTH, cal-

ium, or phosphorous levels, by BMP-7, implicatesMP deficiency in the pathogenesis of renal os-

eodystrophy.

he Function of BMPs in the Kidney

BMPs have emerged (through genetic studies)s important regulators of kidney development.32,33

n addition to being present during kidney devel-pment, BMP-7 messenger RNA also is expressedn the adult kidney, predominantly in the tubules ofhe outer medulla, podocytes, and ureter.34 Further-ore, BMP-7 expression has been shown in cul-

ured kidney cells including glomerular cells andistal Madin Darby kidney cell (MDKC), but notuman proximal HK-2 cells.35

BMP-7 prevented renal failure in a variety ofnimal models of renal injury including acuteenal ischemia, unilateral ureteral obstruction,iabetic nephropathy, and lupus nephritis.36-40 Inddition, treatment with BMP-7 was superior toreatment with enalapril in preventing tubuloin-erstitial fibrosis in the unilateral ureteral ob-truction and lupus nephritis models. Tubuloin-erstitial fibrosis is a major component of severalidney diseases associated with the progression

o CKD stage V. Finally, treatment with BMP-7 c

n the diabetic nephropathy model was moreffective than enalapril in reversing proteinuria,reserving glomerular filtration rate, and pre-enting glomerular sclerosis. Taken together, theMP-7 actions appear to be preservation of ep-

thelial phenotype,40 inhibition of epithelial-esenchymal transdifferentiation,37,40 and inhi-

ition of injury-induced epithelial cellpoptosis.37

MPs in the Vasculature

Vascular calcification is a common problemhat correlates with increased cardiovascularortality among patients with CKD. The process

ppears to be regulated by an osteoblast-likeell,41 possibly of vascular smooth muscle cellVSMC) origin. VSMCs share a common pro-enitor lineage with osteoblasts and retain suf-cient pluripotentiality to transdifferentiate.SMCs are a potential therapeutic target forMP-7 because they have been shown to inhibit

mooth muscle proliferation and to stimulatexpression of markers of smooth muscle cells initro.42 In low-density lipoprotein receptor�/�ice, an animal model of atherosclerosis and

ascular calcification, BMP-7 prevented vascu-ar calcification in uremic animals,43 suggestingpossible protective role for BMP-7 in the vas-

Fig 3. The relationshipbetween secondary hyper-parathyroidism and theosteodystrophy of CKD.Because of decreased os-teoblast differentiation as aresult of CKD, skeletal re-sistance to PTH actions de-velop, leading to changesin phosphorus and calciumhomeostasis and second-ary hyperparathyroidism.The increased PTH levelsrestore rates of bone re-modeling, but an osteodys-trophy develops becausePTH stimulates an abnor-mal phenotype of osteo-blastic cells. Decreasedmineralization (hyperos-teoidosis) and increasedbone resorption result, thelatter caused by PTH stim-ulation of excess RANKLproduction by preosteo-blasts and stromal cells.

ulature among patients with CKD.

tcispal

O

bgnrtbttRbrtd

uta

ORpbmteO

rabcsaobtaotttrrtan

OsbccimiRottrcfhllRoi

RENAL OSTEODYSTROPHY 29

CELL-CELL CONNECTIONS

In states in which there is decreased mineraliza-ion of bone, there is an association with vascularalcification. Various regulatory proteins involvedn bone metabolism and mineralization exist inerum and the vasculature. The role of 2 suchroteins, OPG and �2-HS glycoprotein (AHSG),re discussed with regard to bone and the vascu-ature in patients with CKD.

PG in Bone

With the discovery of OPG, osteoblasts haveeen implicated in the regulation of osteoclasto-enesis (Fig 4). OPG initially was isolated as aovel member of the tumor necrosis factor (TNF)eceptor superfamily made by stromal cells, preos-eoblasts, and preosteoclasts. Unlike most mem-ers of the TNF receptor superfamily, which areransmembrane proteins, OPG is a secreted pro-ein.44 The discovery that OPG can bind toANKL, and compete with RANK, a membrane-ound receptor found on osteoclast precursor cells,evealed a complex mechanism regulating the cell-o-cell contact–dependent process of osteoclastifferentiation.RANKL, present in a membrane-bound or sol-

ble form, is produced by the osteoblast and bindso RANK to stimulate osteoclast formation and

Fig 4. RANKL/RANK/PG regulation of bone re-orption. RANKL producedy bone marrow stromalells, preosteoblasts, andells of the immune system

n either a soluble or cellembrane–anchored form

nteracts with the receptorANK on cells in the oste-clast lineage supportingheir terminal differentiationo multinucleated bone-esorbing cells. OPG is a se-reted circulating receptoror RANKL, making it an in-ibitor of osteoclast stimu-

ation by the RANKL/RANKigand/receptor complex.ANKL is the main regulatorf mature osteoclast activ-

ty.

ctivity (Figs 1 and 4). However, the binding of v

PG to the soluble or membrane-bound form ofANKL inhibits osteoclastogenesis through com-etition of binding and prevention of RANKLinding to RANK on the osteoclast precursor cellembrane. Therefore, it has been suggested that

he degree of osteoclastogenesis and resulting bonerosion is dependent on the ratio of RANKL toPG.45

Studies performed in animal models support theole of OPG and RANKL in osteoclast activationnd differentiation. Mice genetically engineered toe OPG deficient (opg�/opg�) exhibit, by adoles-ence and adulthood, an early onset of osteoporo-is.46 Intravenous injection of recombinant OPGnd transgenic overexpression of OPG in opg�/pg� mice effectively rescued this osteoporoticone phenotype.47 Overexpression of OPG inransgenic mice resulted in osteopetrosis, a state ofbnormally dense bone caused by inhibition ofsteoclast activity.44 In vitro cell culture was usedo determine that OPG inhibits osteoclast matura-ion in a dose-dependent manner.44 In vivo, injec-ion of OPG resulted in osteoclast apoptosis.48 Inats, in vitro studies showed incubation of matureat osteoclasts with RANKL stimulated themhrough multiple cycles of bone resorption.49 Inddition, in vitro studies indicate that RANKL isecessary but not sufficient for osteoclast sur-

48

ival.

A

tmtfoihasdtltntsmprpcci

(bsmsAfwftbmsvAn

ictacii

olSapt

btTBaotbtaFitctTwcifwtTtcAi

fwagmtl(ttcTpf

HRUSKA ET AL30

HSG Glycoprotein in Bone

AHSG is a noncollagenous protein found both inhe serum and in bone matrix. In mineralized hu-an tissues, AHSG has been shown to be concen-

rated with respect to other plasma proteins byactors of 30 to 100.50 However, the concentrationf AHSG in bone and in serum is not constant ands age dependent. The protein appears to be inigher concentrations in children as compared withdults.51,52 In addition, in states of high bone re-orption such as renal osteodystrophy and Paget’sisease, the content of AHSG in bone is higherhan in normal controls.53 However, the serumevels of AHSG in patients with CKD stages Ihrough V and Paget’s disease is lower than inormal controls.54-56 This has led to speculationhat AHSG is involved in bone metabolism. Initialtudies supported this by revealing that AHSGodulates bone resorption in a concentration-de-

endent manner.57 The ability to modulate boneesorption may be related in part to osteoclastrecursors. Osteoclasts are derived from mono-ytes and AHSG has been shown to increase re-ruitment and enhance the function of monocytesn vitro.58-60

AHSG is a 2-chain protein composed of a heavyA) and light (B) chain connected by a disulfideond.61 AHSG has been shown to be structurallyimilar to bovine fetuin, a protein that makes up theajor portion of bovine fetal serum.62,63 This led to

peculation that fetuin is the bovine homologue ofHSG, designating AHSG as fetuin-A or human

etuin. Furthermore, a striking sequence homologyas found between AHSG and the cystatin super-

amily of cysteine proteinase inhibitors.64,65 Fur-her analysis of the A chain of fetuin revealed it toe composed of 5 disulfide loops arranged in aanner similar to the disulfide loops of the cystatin

uperfamily.65 Because cysteine proteinases are in-olved in bone resorption, it was speculated thatHSG may inhibit bone resorption in this man-er.64

The function of AHSG/fetuin has been exam-ned in bone cell cultures. When grown in mediaontaining calcium and phosphorous in concentra-ions that cause spontaneous salt precipitation, theddition of bovine fetuin to rat calvaria osteoblastultures inhibited apatite formation.66 Similar find-ngs occurred when other fetuins were used includ-

ng AHSG. It also was found that the concentration t

f fetuin necessary to prevent precipitation is muchower than the concentration of fetuin in serum.ubsequently, it also was determined that themino acids in the cystatin-like domain 1 of therotein mediated the inhibition of apatite forma-ion.66

Fetuin also has been shown to have effects onone independent of apatite inhibition. Bovine fe-uin has been found to be an antagonist to theGF-� cytokines.67 Fetuin binds to BMP-2 �MP-4 � BMP-6 � TGF-�1 � TGF-�2. Fetuinnd TGF-� receptor type II share sequence homol-gy with a disulfide looped sequence termed TRH1hat is the major cytokine binding domain.67 Theinding of fetuin to the cytokine blocks binding tohe receptor and has been shown to inhibit thectivity of TGF-�1 and BMP-2 in cell culture.67

urther studies have shown that although TGF-�1s required for osteogenesis in dexamethasone-reated rat bone marrow cells, TGF-�1 in highoncentrations inhibits osteogenesis.68 The addi-ion of bovine fetuin to the high concentrationGF-�1 cultures restores osteogenesis. However,hen added to cultures in the presence of low

oncentrations of TGF-�1, bovine fetuin is inhib-tory. The inhibition of osteogenesis by bovineetuin or high-concentration TGF-�1 also occursell before bone mineralization.68 This suggests

hat the relative concentrations of fetuin andGF-�1 play an important role in regulating os-

eogenesis. In addition, this biphasic response alsoan explain how, despite being antagonists, bothHSG/fetuin and BMPs can have protective roles

n vascular calcification (see later).Studies with mice genetically engineered to be

etuin deficient (ahsg�/ahsg�) have had some-hat inconsistent results. One study showed thathsg�/ahsg� mice were fertile and showed noross skeletal abnormalities at day 1 or at 4onths.69 Another study found impaired matura-

ion of growth plate cartilage and slower femurengthening among ahsg�/ahsg� mice later in life3-18 mo).70 Interestingly, however, bone forma-ion was increased, manifested by greater corticalhickness, increased trabecular remodeling, and in-reased osteoblast numbers on bone surfaces.70

he increased bone formation may be caused inart by the absence of the inhibitory effects ofetuin on TGF-�1 or BMPs at physiologic concen-

rations.

O

taePg�secdcmetctctttigcaicrMIlfIssswcw

Rwetmctvs

maRf

erMrmcaevemagmtMccVlailcapdsRvwd

ocrcrfaawvws

RENAL OSTEODYSTROPHY 31

PG in the Vasculature

Apart from bone, OPG is found in a number ofissues including the major arteries such as thebdominal aorta.44 Furthermore, OPG is highlyxpressed in VSMCs, but not endothelial cells.latelet-derived growth factor, basic fibroblastrowth factor, angiotensin II, tumor necrosis factor, and interleukin-1� all up-regulate OPG expres-ion in VSMCs.71 In addition to osteoporosis, micengineered to be opg�/opg� exhibit medial cal-ification of the aorta and renal arteries.46 Theseata support a link between osteoporosis and vas-ular calcification wherein decreased orthotopicineralization leads to increased pressure for het-

rotopic mineralizaton. However, despite the facthat intravenous injection of recombinant OPG res-ued the osteoporosis phenotype, it did not reversehe arterial calcification in opg�/opg� mice.47 Inontrast, transgenic OPG delivered from midges-ation to adulthood does prevent arterial calcifica-ion in opg�/opg� mice,47 suggesting that al-hough OPG can help prevent arterial calcification,t cannot reverse it. Similarly, mice null for matrixla-protein (MGP) develop severe medial vascularalcification and die at age 1 month from coronaryrtery disease and vascular aneurysms.72 This an-mal model has similarities to a rat model of vas-ular calcification induced by warfarin.25,73 Warfa-in inhibits �-carboxylation of the gla residues on

GP and results in aortic calcification in the rat.nterestingly, warfarin therapy of humans does notead to vascular calcification. OPG therapy of war-arin-induced vascular calcification eliminates it.73

n addition, other inhibitors of bone resorptionuch as the bisphosphonates and SB 242784, aelective inhibitor of the osteoclastic V-H-adeno-ine triphosphatase, have been shown to inhibitarfarin- and vitamin D–induced vascular calcifi-

ation.24,74 These findings all support a causal linkith bone mineralization and medial calcification.Although OPG normally is present in arteries,

ANKL and RANK are not detected in the arterialalls of wild-type opg�/opg� adult mice.47 How-

ver, RANKL and RANK transcripts are found inhe calcified arteries of opg�/opg� mice. Further-ore, RANK expression in the calcified arteries

oincides with the presence of multinucleated os-eoclast-like cells. The presence of osteoclast acti-ators and osteoclast-like cells in the calcified ves-

els may imply an attempt to inhibit the m

ineralization by stimulating bone resorption. Thislso further emphasizes the importance of theANKL:OPG ratio in modulating osteoclast dif-

erentiation and bone resorption.Studies of human medial calcification and ath-

rosclerosis have revealed possible roles for boneegulating proteins in vascular calcification. With

onckeberg’s sclerosis, calcification occurs in di-ect apposition to VSMCs without the presence ofacrophages in lipids seen in the intimal calcifi-

ation of atherosclerosis.75 These VSMCs directlybutting the area of calcification express high lev-ls of MGP. In contrast, as compared with normalessels, vessels with medial calcification globallyxpress lower levels of MGP and higher levels ofarkers for osteoblasts and chondrocytes such as

lkaline phosphatase, bone sialoprotein, and colla-en II.75 The low levels of MGP in vessels withedial calcification may suggest a predisposition

o calcification. The localized up-regulation ofGP by VSMCs may reflect a reaction to the

alcification and an attempt to increase calciumlearance. This is supported by the fact that whenSMCs calcify, MGP expression is up-regu-

ated.76 Similarly, in a study of human calcifiedtherosclerotic plaques, expression of calcificationnhibitors such as MGP and OPG are down-regu-ated whereas markers of osteoblasts and chondro-ytes were up-regulated.77 Another study of humantherosclerotic plaques also found increased ex-ression of MGP along the boundary of mineraleposits in advanced calcific lesions similar to thateen with purely medial calcification.78 In addition,ANKL, weakly expressed in VSMCs in normalessels, only could have been shown in associationith the extracellular matrix surrounding calciumeposits.OPG also has been studied as a potential marker

f cardiovascular disease. In one study, 490 Cau-asian women of at least 65 years of age wereecruited to assess whether OPG levels were asso-iated with stroke, mortality, and cardiovascularisk factors such as diabetes mellitus. The studyound that OPG levels obtained at baseline werebout 30% greater in women with diabetes mellitusnd that increased OPG levels were associatedith an increase in all-cause mortality and cardio-ascular mortality.79 OPG levels did not correlateith C-reactive protein levels in these patients,

uggesting that OPG was not a marker of inflam-

ation. Another study evaluated whether OPG was

atgllsoiOmtbcccipcRctaiRgcplwdsRi

aulpcstuHVtiilsst

sbsbt(ninsstnbsolefifiapieus

A

isatetMeipyrmposcbMri

HRUSKA ET AL32

ssociated with the presence of coronary disease. Aotal of 201 patients who underwent coronary an-iography because of chest pain had serum OPGevels measured. The study found that serum OPGevels were increased significantly in patients withignificant stenoses as compared with those with-ut stenoses. In addition, as the severity of diseasencreased, there also was a subsequent increase inPG levels.80 Another study with 522 Caucasianen undergoing coronary angiography showed

hat once again there was a positive correlationetween serum OPG levels and the severity oforonary disease.81 In addition, OPG levels in-reased with age and were increased in diabetic asompared with nondiabetic patients. Finally, thenvestigators of the previous study also have re-orted that among 346 Caucasian men undergoingoronary angiography, serum levels of solubleANKL were significantly lower in patients withoronary disease than those without. However,here was no correlation between soluble RANKLnd the severity of coronary disease.82 This is annteresting finding because it suggests theANKL:OPG ratio in diseased vessels is muchreater than that found in serum among those withardiovascular disease. How this applies to theathogenesis of cardiovascular disease and vascu-ar calcification, particularly in CKD patients inhom OPG levels accumulate in part owing toecreased clearance (see later), remains to beeen. Clearly, however, it again appears that theANKL:OPG ratio is more important than the

ndividual concentrations.Why OPG is down-regulated in diseased tissues

nd up-regulated in the serum in vascular disease isnclear. Animal models suggest that low OPGevels predispose to vascular calcification. Theresence of RANKL, RANK, and osteoclast-likeells in the calcified vessels of opg�/opg� miceuggests a protective mechanism against calcifica-ion by osteoclastogenesis. Perhaps the down-reg-lation of OPG also plays a role in this manner.owever, it is possible that the role of OPG inSMCs is independent and separate from regula-

ion of osteoclastogenesis. TNF-related apoptosis-nducing ligand (TRAIL) plays an important rolen a variety of processes including inducing cellu-ar apoptosis. TRAIL is found in a variety of tis-ues throughout the body including vascularmooth muscle and has been found to be cytotoxic

o both vascular endothelial cells and medial m

mooth muscle cells.83 OPG has been found toind TRAIL and inhibits TRAIL-induced apopto-is.84 In addition, OPG expression in VSMCs haseen shown to be down-regulated by ligands forhe peroxisome proliferator-activated receptor-�PPAR-�), part of a family of ligand-activateduclear transcription factors. PPAR-� has beendentified as a nuclear receptor for thiazolidinedio-es, compounds used to treat diabetes as insulinensitizers.85 Ligands for PPAR-� have beenhown to inhibit VSMC proliferation86 and thus arehought to have anti-atherogenic properties. Fi-ally, platelet-derived growth factor and basic fi-roblast growth factor, known to stimulate expres-ion of OPG in VSMCs, also are potent stimulatorsf VSMC proliferation and migration. Moreover,igands for PPAR-� have been shown to inhibit theffects of platelet-derived growth factor and basicbroblast growth factor on VSMCs.87 All of thesendings taken together suggest that OPG may playregulatory role in VSMCs. Decreased OPG ex-

ression in diseased vessels may be an attempt tonhibit VSMC proliferation and migration. How-ver, further studies need to be performed, partic-larly in regard to what role, if any, the increasederum OPG plays in the pathogenesis.

HSG in the Vasculature

In addition to osteoblast cell cultures, fetuinnhibits apatite formation in vitro, suggesting aimilar role for fetuin in the serum.66 Recently, inn attempt to understand how MGP inhibits ec-opic calcification, a new circulating protein min-ral complex was discovered. The composition ofhis complex consists of mineral, fetuin, and

GP.88 The protein mineral complex was discov-red after rats were injected with etidronate, result-ng in the increase of nonionic calcium, as well ashosphorous and MGP, in serum. Subsequent anal-sis determined the nonionic calcium, phospho-ous, and MGP to be components of this proteinineral complex. Fetuin was discovered to be the

redominant molecule in this complex. The sourcef the fetuin content was found to be derived fromerum, consuming nearly half of the serum fetuinontent without affecting total serum levels (50%ound to the protein mineral complex, 50% free).oreover, the protein mineral complex is cleared

apidly from blood by 9 to 24 hours after etidronatenjection. The investigators concluded that the for-

ation of the complex occurred as a result of

ibtbabitjrssqp

maTbcpeascm5iatccmptfcciMtntmw

mtclf

pwtmdta

etfcsiobg

pnhccmsfyvds

pfAiokiAratsimptcat

RENAL OSTEODYSTROPHY 33

nhibition of bone mineralization rather than inhi-ition of bone resorption. This was supported byhe fact that the appearance of the complex occursefore any expected inhibition of bone resorptionnd by the fact that treatment with alendronate, aisphosphonate that did not inhibit bone mineral-zation, did not result in the formation of the pro-ein mineral complex.88 In addition, etidronate in-ection in rats pretreated with warfarin did notesult in an increase in serum MGP levels. Thisuggests that the increase in MGP is caused by newynthesis and the �-carboxylation of MGP is re-uired for its binding to the protein mineral com-lex.In a second study, the formation of the fetuinineral complex was inhibited by calcitonin, OPG,

nd alendronate—all inhibitors of bone resorption.he investigators suggested that this implied aone origin of the mineral complex.89 This wasorroborated by a similar increase in calcium,hosphorous, and MGP levels after injection withtidronate among mice fed a calcium-deficient diets compared with those fed a calcium-replete diet,uggesting bone as the origin of the increasedalcium. A third study found that after the proteinineral complex was cleared, there was a nearly

0% reduction of the serum fetuin level, suggest-ng that clearance of the protein mineral complexlso cleared the associated fetuin.90 In addition, thehird study also showed that the protein mineralomplex also contained an additional protein, se-reted phosphoprotein 24, a protein similar in do-ain structure to fetuin. Exogenous secreted phos-

hoprotein 24 was found to associate strongly withhe protein mineral complex when added to it. Aourth study documented that the addition of cal-ium and phosphate to rat serum, increasing theoncentration of each by 10 nmol/L, also resultedn the formation of this protein mineral complex.91

oreover, bovine fetuin inhibits mineral precipi-ation when added to solutions containing 5mol/L of phosphate and calcium, and this inhibi-ion is associated with the formation of a fetuin-ineral complex. This fetuin-mineral complex alsoas found to bind MGP strongly.Finally, a recent study showed, using electronicroscopy and dynamic light scattering, that apa-

ite precipitation inhibition by fetuin/AHSG isaused by the transient formation of soluble, col-oid spheres containing calcium, phosphorous, and

92

etuin/AHSG in vitro. The findings also support t

revious reports that the cystatin-like domain 1as critical to precipitation inhibition. The inves-

igators also speculated that these colloid spheresay be identical to the protein mineral complex

escribed earlier. These findings all suggest thathis circulating protein mineral complex serves asn inhibitor of mineral precipitation in serum.

Among ahsg�/ahsg� mice, some developedctopic microcalcifications in soft tissues.69 Also,he addition of BMPs to ectopic sites resulted inurther bone extension in ahsg�/ahsg� mice asompared with ahsg�/ahsg� mice.70 This furtherupports the concept that fetuin/AHSG acts as annhibitor of mineralization. Among calcificationsf atherosclerotic human aortic tissue, AHSG haseen found in concentrations approximately 7-foldreater than that in plasma.93

As stated earlier, AHSG levels are lower amongatients with CKD stages I through V than withormal controls. In patients with CKD stage V onemodialysis, low concentration of AHSG is asso-iated with increased C-reactive protein, enhancedardiovascular mortality, and enhanced all-causeortality.56 In addition, the sera of these patients

howed impaired ex vivo ability to inhibit apatiteormation. Patients on long-term hemodialysis (�2) showed a trend toward increased severity ofascular calcification than patients on short-termialysis (�1 y), however, this was not statisticallyignificant.

Another study found that nondiabetic patientsresenting with acute ST elevation myocardial in-arction showed a trend toward lower serumHSG levels than normal controls.94 AHSG levels

n these patients also trended back to normal levelsn discharge. This finding suggests the long-nown idea that AHSG is also a negative marker ofnflammation. Previous studies have shown thatHSG has a negative correlation with acute phase

eactants such as �1-antitrypsin and haptoglobinnd a positive correlation with albumin.95 In addi-ion, interleukin 1� and interleukin 6 have beenhown to down-regulate expression of AHSG andts messenger RNA in cell culture similar to albu-in.96 Because inflammation is thought to play a

ivotal role in the pathogenesis of atherosclerosis,he association of low AHSG levels and cardiovas-ular mortality among patients with CKD stage Vnd nondiabetic patients presenting with ST eleva-ion myocardial infarction, may in part be related

o inflammation.

O

attpowphstpmdwv

powthsatt(rmiphEaIiO7cidt

s3odmp

temaemrdPrtrprpdto

elceootpdsmf

A

laAchptCmcidahma

HRUSKA ET AL34

PG in Patients With CKD

Two recent reports indicate that levels of OPGre increased in patients with CKD stages Ihrough V.97,98 One of these reports98 determinedhat OPG levels increase with age in a healthyopulation as well as in patients with CKD stage Vn hemodialysis. The levels in the latter, however,ere significantly higher independent of age. Inatients with CKD stages I through V not onemodialysis, serum levels of OPG correlated witherum creatinine levels and had a reciprocal rela-ionship to creatinine clearance over a 24-houreriod. Thus, the kidney was reported to be theajor site for clearance of OPG. In addition, it was

etermined that OPG found in serum of patientsith CKD was capable of binding to RANKL initro, showing putative bioactivity in vivo.Recent studies have looked at serum OPG as a

ossible predictor of bone turnover rate in renalsteodystrophy. One study examined 26 patientsith CKD stage V just before kidney transplanta-

ion.99 These patients had been on maintenanceemodialysis from 9 to 67 months at the time ofampling. Determination of bone mineral densitynd bone histomorphometry allowed the correla-ion of bone morphology to markers of bone me-abolism. The patients were divided into type IInormal or low turnover) and III (high turnover)enal osteodystrophy based on the histomorpho-etric results. In corroboration of previous studies

t was determined that significantly high intactarathyroid hormone (iPTH) levels correlated withistomorphometric indices of high bone turnover.xamination of OPG indicated increased levels inll dialysis patients, but with lower amounts in typeII as compared with type II disease state. Finally,t was reported that the combination of iPTH andPG levels correctly classified types II and III in2% and 88% of the patients, respectively. Thusoncluding that measuring circulating OPG andPTH levels might be a valid noninvasive means ofetermining bone turnover rate in renal osteodys-rophy.

A second study also examined bone biopsypecimens along with markers of bone turnover in9 patients with CKD stage V.100 The average timen dialysis was 60.9 � 83.4 months. Patients wereescribed as having adynamic bone disease, osteo-alacia (low turnover states), predominant hyper-

arathyroidism, or mixed osteodystrophy (high A

urnover states). Results similar to those reportedarlier indicated that OPG levels were increasedarkedly over the normal range in both the low

nd high turnover conditions. Correlation of dis-ase state with OPG levels in conjunction with thearkers of bone remodeling indicated significant

elationships. Similar to what was reported earlier,iagnosis of high turnover osteodystrophy, andTH levels less than 1,000 pg/mL, OPG was cor-elated inversely to iPTH, total PTH, and parame-ers of bone resorption. In addition, negative cor-elations were reported between OPG andarameters of bone formation. However, when pa-ameters of patients with iPTH levels less than 300g/mL were examined, those with adynamic boneisease were found to have lower OPG levels thanhose diagnosed with either hyperparathyroidismr mixed osteodystrophy.The discrepancy between the 2 studies described

arlier, in which attempts were made to use OPGevels to predict degree of renal osteodystrophy,ould be owing to the small number of patientsxamined. This incongruity also could reflect thatther factors, as noted earlier, are involved in thesteoblast regulation of osteoclastogenesis. The ra-io of RANKL to OPG is actually the importantarameter in determining the progress of osteoclastevelopment and activation in the bone, thus mea-uring the serum levels of OPG rather than deter-ining ratios of RANKL to OPG, a more difficult

eat, appears to have significant shortcomings.

HSG in Patients With CKD

As stated earlier, AHSG levels appear to beower among patients with CKD stages I through Vs compared with normal controls. In addition,HSG appears to have increased expression in

alcified vessels. This is in contrast to OPG, whichas increased serum expression and decreased ex-ression in calcified vessels. Possible causes forhe decreased serum AHSG levels in patients withKD include chronic inflammation and proteinalnutrition. Current theories regarding vascular

alcification suggest that decreased bone mineral-zation leads to heterotopic mineralization. In rats,ecreased mineralization leads to the formation ofprotein mineral complex containing fetuin, the ratomologue of AHSG. The purpose of this proteinineral complex is thought to be inhibition of

patite precipitation at ectopic sites. Low levels of

HSG in these patients also predicted an increased

cariwiscwV

abtApBtttiigmlcbpbBv

cpalsrhlmtnammrl

pc

bbavgpebesctOtAciesacp

m2

r

aa1

mp

d1

t2

p2

k

BbP

RENAL OSTEODYSTROPHY 35

ardiovascular risk. It certainly is conceivable thatmong patients with CKD, low levels of AHSGesult in a reduced ability to inhibit apatite precip-tation. Indeed, it was documented that patientsith CKD stage V on hemodialysis had such an

mpaired inhibition. In addition, fetuin has beenhown to modulate osteogenesis. What role de-reased levels of AHSG in CKD patients playsith regard to osteogenesis, either in bone or inSMCs, remains to be seen.

CONCLUSION

The relationship between bone and the kidney iscomplex interplay of kidney-bone connections,

one-kidney connections, and cell-cell connec-ions. Bone is in a constant state of remodeling.lterations in remodeling rates lead to adaptiverocesses to correct it. In CKD, deficiency ofMP-7 leads to decreased osteoblast differentia-

ion and decreased bone remodeling. As an adap-ive mechanism, PTH levels increase in an attempto restore normal bone remodeling. However, PTHs not an osteoblastic differentiation factor. Instead,ncreased levels of PTH stimulate osteoblast pro-enitors to become fibroblast-like cells, resulting inarrow fibrosis. In addition, PTH stimulation

eads to increased RANKL expression and de-reased OPG, resulting in osteoclast formation andone resorption. Treatment of this maladaptiverocess results in quiescent bone or the adynamicone disorder. We have shown that treatment withMP-7 in animal models of CKD effectively re-erses both of these high and low turnover states.Second, vascular calcification is an important

ause of morbidity and mortality in CKD. Theathogenesis of medial calcification likely involvesn osteoblast-like cell within the vessel wall. Theikely candidate for this is the VSMC because ithares common progenitors with osteoblasts andetains the ability to transdifferentiate. Evidenceas shown that decreased bone mineralizationeads to increased pressure toward heterotopicineralization. BMP-7 has been shown to induce

he phenotype of VSMCs. In addition, by restoringormal bone mineralization, BMP-7, in theory,lso decreases the pressure toward heterotopicineralization. Indeed, we have shown that treat-ent with BMP-7 in animal models of atheroscle-

osis and vascular disease in CKD reverses vascu-

ar calcification. Thus, it appears that BMP-7 is a b

otential therapy for renal osteodystrophy and vas-ular calcification.

Finally, 2 bone regulating proteins found inone, serum, and the vasculature also play a role inone remodeling and vascular calcification. OPGnd AHSG, the human homologue of fetuin, havearying roles in these processes. Indeed, mice en-ineered to be OPG deficient show severe osteo-orosis and vascular calcification, whereas thosengineered to be fetuin deficient have increasedone formation and an increased susceptibility toctopic calcification. Among patients with CKD,erum OPG levels appear higher than in normalontrols whereas serum AHSG levels are lowerhan normal in controls. Higher serum levels ofPG correlate with increased cardiovascular mor-

ality, diabetes, and coronary disease whereas lowHSG levels correlate with increased cardiovas-

ular mortality. AHSG modulates osteogenesis andnhibits apatite formation by forming soluble min-ral complexes, indicating a protective role of thisubstance against vascular calcification. The over-ll role of OPG in vascular calcification is lesslear, and likely the ratio of RANKL and OPGlays a more important role.

REFERENCES

1. Elder G: Pathophysiology and recent advances in theanagement of renal osteodystrophy. J Bone Miner Res 17:

094-2105, 20022. Ho LT, Sprague SM: Renal osteodystrophy in chronic

enal failure. Semin Nephrol 22:488-493, 20023. Goodman WG, Ramirez JA, Belin T: Development of

dynamic bone in patients with secondary hyperparathyroidismfter intermittent calcitriol therapy. Kidney Int 46:1160-1166,9944. Lund RJ, Davies MR, Brown AJ, et al: Successful treat-ent of an adynamic bone disorder with bone morphogenetic

rotein-7 in a renal ablation model. J Am Soc Nephrol (in press)5. Reddi AH: Bone morphogenetic proteins and skeletal

evelopment: The kidney-bone connection. Pediatr Nephrol4:598-601, 20006. Groeneveld EHJ, Burger EH: Bone morphogenetic pro-

eins in human bone regeneration. Eur J Endocrinol 142:9-21,0007. Miyazono K: Signal transduction by bone morphogenetic

rotein receptors: Functional roles of Smad proteins. Bone5:91-93, 19998. Miyazono K: TGF-� signaling by Smad proteins. Cyto-

ine Growth Factor Rev 11:15-22, 20009. Yamaguchi A, Ishizuya T, Kintou N, et al: Effects of

MP-2, BMP-4, and BMP-6 on osteoblastic differentiation ofone marrow-derived stromal cell lines, ST2 and MC3T3-G2/A6. Biochem Biophys Res Commun 220:366-371, 199610. Hughes FJ, Collyer J, Stanfield M, et al: The effects of

one morphogenetic protein-2, -4, and -6 on differentiation of

r1

bd

mhb1

bfnl2

og1

tt

mdf

a

tg2

ga

g

Bmp

ba

dJ

P

tP

bta7

m

t1

hCm

ooMf

ep1

me

AaG

g9

bm

tP

tj1

ts

no

pK

yei

nn

nca1

ca1

HRUSKA ET AL36

at osteoblast cells in vitro. Endocrinology 136:2671-2677,99511. Ebisawa T, Tada K, Kitajima I, et al: Characterization of

one morphogenetic protein-6 signaling pathways in osteoblastifferentiation. J Cell Sci 112:3519-3527, 199912. Lecanda F, Avioli LV, Cheng SL: Regulation of boneatrix protein expression and induction of differentiation of

uman osteoblasts and human bone marrow stromal cells byone morphogenetic protein-2. J Cell Biochem 67:386-398,99713. Sampath TK, Maliakal JC, Hauschka PV, et al: Recom-

inant human osteogenic protein-1 (hOP-1) induces new boneormation in vivo with a specific activity comparable withatural bovine osteogenic protein and stimulates osteoblast pro-iferation and differentiation in vitro. J Biol Chem 267:20352-0362, 199214. Gruber R, Mayer C, Schulz W, et al: Stimulatory effects

f cartilage-derived morphogenetic proteins 1 and 2 on osteo-enic differentiation of bone marrow stromal cells. Cytokine2:1630-1638, 200015. Chaudhary LR, Hofmeister AM, Hruska KA: Differen-

ial growth factor regulation of bone formation by human os-eoprogenitor cells in vivo. Bone (in press)

16. Amedee J, Bareille R, Rouais F, et al: Osteogenin (boneorphogenic protein 3) inhibits proliferation and stimulates

ifferentiation of osteoprogenitors in human bone marrow. Dif-erentiation 58:157-164, 1994

17. Bahamonde ME, Lyons KM: BMP3: To be or not to beBMP. J Bone Joint Surg 83:S1-56-S1-62, 200118. Faucheux C, Ulysse F, Bareille R, et al: Opposing ac-

ions of BMP3 and TGF�1 in human bone marrow stromal cellrowth and differentiation. Biochem Biophys Res Commun41:787-793, 199719. Jane JA Jr, Dunford BA, Kron A, et al: Ectopic osteo-

enesis using adenoviral bone morphogenetic protein (BMP)-4nd BMP-6 gene transfer. Mol Ther 6:464-470, 2002

20. Alden TD, Varady P, Kallmes DF, et al: Bone morpho-enetic protein gene therapy. Spine 27:587-593, 200321. Nochi H, Sung JH, Lou J, et al: Adenovirus mediated

MP-13 gene transfer induces chondrogenic differentiation ofurine mesenchymal progenitor cells. J Bone Miner Res (in

ress)22. Simon M, Maresh JG, Harris SE, et al: Expression of

one morphogenetic protein-7 mRNA in normal and ischemicdult rat kidney. Am J Physiol 276:F382-F389, 1999

23. Wang S, Hirschberg R: Loss of renal tubular BMP7uring the evolution of experimental diabetic nephropathy.Am Soc Nephrol 11:655A, 2000 (abstr)24. Hruska KA: Pathophysiology of renal osteodystrophy.

ediatr Nephrol 14:636-640, 200025. Hruska KA: Growth factors and cytokines in renal os-

eodystrophy, in Bushinsky DA (ed): Renal Osteodystrophy,hiladelphia, 1998, pp 221-26126. Nishida S, Yamaguchi A, Tanizawa T, et al: Increased

one formation by intermittent parathyroid hormone adminis-ration is due to the stimulation of proliferation and differenti-tion of osteoprogenitor cells in bone marrow. Bone 15:717-23, 199427. Isogai Y, Akatsu T, Ishizuya T, et al: Parathyroid hor-

one regulates osteoblast differentiation positively or nega-

ively depending on the differentiation stages. J Bone Miner Res1:1384-1393, 199628. Bogdanovic Z, Huang Y-F, Dodig M, et al: Parathyroid

ormone inhibits collagen synthesis and the activity of ratollal transgenes mainly by a cAMP-mediated pathway inouse calvariae. J Cell Biochem 77:149-158, 200029. Chaudhary LR, Avioli LV: Identification and activation

f mitogen-activated protein (MAP) kinase in normal humansteoblastic and bone marrow stromal cells: Attenuation ofAP kinase activation by cAMP, parathyroid hormone and

orskolin. Mol Cell Biochem 178:59-68, 199830. Uzawa T, Hori M, Ejiri S, et al: Comparison of the

ffects of intermittent and continuous administration of humanarathyroid hormone (1-34) on rat bone. Bone 16:477-484,99531. Gonzalez EA, Lund RJ, Martin KJ, et al: Treatment of aurine model of high-turnover renal osteodystrophy by exog-

nous BMP-7. Kidney Int 61:1322-1331, 200232. Luo G, Hofmann C, Bronckers AL, Sohocki M, Bradley

, Karsenty G: BMP-7 is an inducer of nephrogenesis, and islso required for eye development and skeletal patterning.enes Dev 9:2808-2820, 199533. Martinez G, Bertram JF.Organization of bone morpho-

enetic proteins in renal development. Nephron Exp Nephrol3:e18-22, 200334. Dudley AT, Lyons KM, Robertson EJ: A requirement for

one morphogenetic protein-7 during development of the mam-alian kidney and eye. Genes Dev 9:2795-2807, 199535. Kitten AM, Kreisberg JI, Olson MS: Expression of os-

eogenic protein-1 mRNA in cultured kidney cells. J Cellhysiol 181:410-415, 199936. Vukicevic S, Basic V, Rogic D, et al: Osteogenic pro-

ein-1 (bone morphogenetic protein-7) reduces severity of in-ury after ischemic acute renal failure in rat. J Clin Invest02:202-214, 199837. Hruska K, Guo G, Wozniak M, et al: Osteogenic pro-

ein-1 prevents renal fibrogenesis associated with ureteral ob-truction. Am J Physiol 279:F130-F143, 2000

38. Morrissey J, Hruska K, Guo G, et al: Bone morphoge-etic protein-7 improves renal fibrosis and accelerates the returnf renal function. J Am Soc Nephrol 13:S14-S21, 200239. Wang S, Chen Q, Simon TC, et al: Bone morphogenetic

rotein-7 (BMP-7), a novel therapy for diabetic nephropathy.idney Int 63:2037-2049, 200340. Zeisberg M, Hanai J, Sugimoto H, Mammoto T, Char-

tan D, Strutz F, Kalluri R: BMP-7 counteracts TGF�1-inducedpithelial-to-mesenchymal transition and reverses chronic renalnjury. Nature Med 9:964-968, 2003

41. Davies MR, Hruska KA: Pathophysiological mecha-isms of vascular calcification in end-stage renal disease. Kid-ey Int 60:472-479, 200142. Dorai H, Vukicevic S, Sampath TK: Bone morphoge-

etic protein-7 (osteogenic protein-1) inhibits smooth muscleell proliferation and stimulates the expression of markers thatre characteristic of SMC phenotype in vitro. J Cell Physiol84:37-45, 200043. Davies MR, Lund RJ, Hruska KA: BMP-7 is an effica-

ious treatment of vascular calcification in a murine model oftherosclerosis and chronic renal failure. J Am Soc Nephrol4:1559-1567, 2003

44. Simonet WS, Lacey DL, Dunstan CR, et al: Osteopro-

tb

ou

dc

vvo

mP

oJ

r2

c7

spp2

t3

c1

s5

ovs

st

oJ

c1

p4

dtB

TF

cJ

d2

rJ

av2

pI1

o

a2

gmg

ts

cp

iD

sa

lwm1

cle

ddA

es2

ssn

RENAL OSTEODYSTROPHY 37

egerin: A novel secreted protein involved in the regulation ofone density. Cell 89:309-319, 199745. Hofbauer LC, Khosla S, Dunstan CR, et al: The roles of

steoprotegerin and osteoprotegerin ligand in the paracrine reg-lation of bone resorption. J Bone Miner Res 15:2-12, 200046. Bucay N, Sarosi I, Dunstan CR, et al: Osteoprotegerin-

eficient mice develop early onset osteoporosis and arterialalcification. Genes Dev 12:1260-1268, 1998

47. Min H, Morony S, Sarosi I, et al: Osteoprotegerin re-erses osteoporosis by inhibiting endosteal osteoclasts and pre-ents vascular calcification by blocking a process resemblingsteoclastogenesis. J Exp Med 192:463-474, 200048. Lacey DL, Tan HL, Lu J, et al: Osteoprotegerin ligandodulates murine osteoclast survival in vitro and in vivo. Am Jathol 157:435-448, 200049. Burgess TL, Qian Y-X, Kaufman S, et al: The ligand for

steoprotegerin (OPGL) directly activates mature osteoclasts.Cell Biol 145:527-538, 199950. Triffitt JT: Plasma proteins in human cortical bone: En-

ichment of the alpha2HS-glycoprotein. Calcif Tissue Int 22:7-33, 197651. Dickson IR, Bagga MK: Changes with age in the non-

ollagenous proteins of human bone. Connect Tissue Res 14:7-85, 198552. Dickson IR, Bagga MK, Patterson CR: Variations in the

erum concentration and urine excretion of alpha 2HS-glyco-rotein, a bone related protein, in normal individuals and inatients with osteogenesis imperfecta. Calcif Tissue Int 35:16-0, 198353. Quelch KJ, Cole WG, Melick RA: Noncollagenous pro-

eins in normal and pathological human bone. Calcif Tissue Int6:545-549, 198454. Ashton BA, Smith R: Plasma alpha 2HS-glycoprotein

oncentration in Paget’s disease of bone. Clin Sci 58:435-438,98055. Kishore BK, Gejyo AM: Alpha 2HS-glycoprotein in the

erum and urine of patients with renal diseases. Postgrad Med9:304-307, 198356. Ketteler M, Bongartz P, Westenfeld R, et al: Association

f low fetuin-A (AHSG) concentrations in serum with cardio-ascular mortality in patients on dialysis: A cross-sectionaltudy. Lancet 361:827-833, 2003

57. Colclasure GC, Lloyd WS, Lamkin M, et al: Humanerum alpha 2HS-glycoprotein modulates in vitro bone resorp-ion. J Clin Endocrinol Metab 66:187-192, 1988

58. Malone JD, Teitelbaum SL, Griffin GL, et al: Recruit ofsteoclast precursors by purified bone matrix constituents.Cell Biol 92:227-230, 198259. Malone JD, Richards M: Alpha 2HS glycoprotein is

hemotactic for mononuclear phagocytes. J Cell Physiol 132:18-124, 198760. Lewis JG, Andre CM: Enhancement of human monocyte

hagocytic function by alpha 2HS glycoprotein. Immunology2:481-487, 198161. Araki T, Yoshioka Y, Schmid K: The position of the

isulfide bonds in human plasma alpha 2HS-glycoprotein andhe repeating double disulfide bonds in the domain structure.iochim Biophys Acta 994:195-199, 198962. Christie DL, Dziegielewska KM, Hill RM, et al: Fetuin:

he bovine homologue of human alpha 2HS glycoprotein.

EBS Lett 214:45-49, 1987

63. Dziegielewska KM, Brown WM, Casey SJ, et al: Theomplete cDNA and amino acid sequence of bovine fetuin.Biol Chem 265:4354-4357, 199064. Elzanowski A, Barker WC, Hunt LT, et al: Cystatin

omains in alpha-2-HS glycoprotein and fetuin. FEBS Lett27:167-170, 198865. Kellerman J, Haupt H, Auerswald EA, et al: The ar-

angement of disulfide loops in human alpha2-HS glycoprotein.Biol Chem 264:14121-14128, 198966. Schinke T, Amendt C, Trindl A, et al: The serum protein

lpha2-HS glycoprotein/fetuin inhibits apatite formation initro and in mineralizing calvaria cells. J Biol Chem 271:20789-0796, 199667. Demetriou M, Binkert C, Sukhu B, et al: Fetuin/al-

ha2-HS glycoprotein is transforming growth factor-beta typeI receptor mimic and cytokine antagonist. J Biol Chem 271:2755-12761, 199668. Binkert C, Demetriou M, Sukhu B, et al: Regulation of

steogenesis by Fetuin. J Biol Chem 274:28514-28520, 199969. Jahnen-Dechent W, Schinke T, Trindl A, et al: Cloning

nd targeted deletion of the mouse fetuin gene. J Biol Chem72:31496-31503, 199770. Szweras M, Lio D, Partridge EA, et al: Alpha2-HS

lycoprotein/fetuin, a transforming growth factor-beta/boneorphogenetic protein antagonist, regulates postnatal bone

rowth and remodeling. J Biol Chem 277:19991-19997, 200271. Zhang J, Fu M, Myles D, et al: PDGF induces osteopro-

egerin expression in vascular smooth muscle cells by multipleignal pathways. FEBS Lett 521:180-184, 2002

72. Luo G, Ducy P, McKee MD, et al: Spontaneous calcifi-ation of arteries and cartilage in mice lacking matrix GLArotein. Nature 386:78-81, 199773. Price PA, June HH, Buckley JR, et al: Osteoprotegerin

nhibits artery calcification induced by warfarin and by vitamin. Arterioscler Thromb Vasc Biol 21:1610-1616, 200174. Price PA, June HH, Buckley JR, et al: SB 242784, a

elective inhibitor of the osteoclastic V-H�-ATPase, inhibitsrterial calcification in the rat. Circ Res 91:547-552, 2002

75. Shanahan CM, Cary NRB, Salisbury JR, et al: Medialocalization of mineralization-regulating proteins in associationith monckeberg’s sclerosis: Evidence for smooth muscle cell-ediated vascular calcification. Circulation 100:2168-2176,

99976. Proudfoot D, Skepper JN, Shanahan CM, et al: Calcifi-

ation of human vascular cells in vitro is correlated with highevels of matrix Gla protein and low levels of osteopontinxpression. Arterioscler Thromb Vasc Biol 18:379-388, 1998

77. Tyson KL, Reynolds JL, McNair R, et al: Osteo/chon-rocytic transcription factors and their target genes exhibitistinct patterns of expression in human arterial calcification.rterioscler Thromb Vasc Biol 23:489-494, 200378. Dhore CR, Cleutjens J, Lutgens E, et al: Differential

xpression of bone matrix regulatory proteins in human athero-clerotic plaques. Arterioscler Thromb Vasc Biol 21:1998-003, 200179. Browner WS, Lui LY, Cummings SR: Associations of

erum osteoprotegerin levels with diabetes, stroke, bone den-ity, fractures, and mortality in elderly women. J Clin Endocri-ol Metab 86:631-637, 2001

80. Jono S, Ikari Y, Shioi A, et al: Serum osteoprotegerin

la

oe

lo

i2

g2

asR

mr

vJ

aaJ

ttc

cC

pf

oJ

ofi5

amC

tmn

ah

l4

cpD

att

t2

HRUSKA ET AL38

evels are associated with the presence and severity of coronaryrtery disease. Circulation 106:1192-1194, 2002

81. Schoppet M, Sattler AM, Schaefer JR, et al: Increasedsteoprotegerin serum levels in men with coronary artery dis-ase. J Clin Endocrinol Metab 88:1024-1028, 2003

82. Schoppet M, Schaefer JR, Hofbauer LC: Low serumevels of soluble RANK ligand are associated with the presencef coronary artery disease in men. Circulation 107:76e, 200383. Gochuico BR, Zhang J, Ma BY, et al: TRAIL expression

n vascular smooth muscle. Am J Physiol 278:L1045-L1050,00084. Emery JG, McDonnell P, Burke MB, et al: Osteoprote-

erin is a receptor for the cytotoxic ligand TRAIL. J Biol Chem73:14363-14367, 199885. Fu M, Zhang J, Lin Y, et al: Activation of peroxisome-

ctivated receptor gamma inhibits osteoprotegerin gene expres-ion in human aortic smooth muscle cells. Biochem Biophyses Commun 294:597-601, 200286. Bishop-Bailey D, Hla T, Warner TD: Intimal smoothuscle cells as a target for peroxisome proliferator-activated

eceptor-gamma ligand therapy. Circ Res 91:210-217, 200287. Law RE, Meehan WP, Xi XP, et al: Troglitazone inhibits

ascular smooth muscle cell growth and intimal hyperplasia.Clin Invest 98:1897-1905, 199688. Price PA, Thomas GR, Pardini AW, et al: Discovery of

high molecular weight complex of calcium, phosphate, fetuin,nd matrix Gla protein in the serum of etidronate-treated rats.Biol Chem 277:3926-3934, 200289. Price PA, Caputo JM, Williamson MK: Bone origin of

he serum complex of calcium, phosphate, and matrix gla pro-ein: Biochemical evidence for the cancellous bone-remodelingompartment. J Bone Miner Res 17:1171-1179, 2002

90. Price PA, Nguyen TMT, Williamson MK: Biochemicalharacterization of the serum fetuin-mineral complex. J Biolhem 278:22153-22160, 2003

91. Price PA, Lim JE: The inhibition of calcium phosphaterecipitation by fetuin is accompanied by the formation of aetuin-mineral complex. J Biol Chem 278:22144-22152, 2003

92. Heiss A, DuChesne A, Denecke B, et al: Structural basisf calcification inhibition by �2-HS glycoprotein/Fetuin-A.Biol Chem 278:1333-13341, 200393. Keeley FW, Sitarz EE: Identification and quantification

f alpha 2-HS glycoprotein in the mineralized matrix of calci-ed plaques of atherosclerotic human aorta. Atherosclerosis5:63-69, 198594. Mathews ST, Deutsch DDIG, Hora N, et al: Plasma

lpha2-HS glycoprotein concentrations in patients with acuteyocardial infarction quantified by a modified ELISA. Clinhim Acta 319:27-34, 200295. Lebreton JP, Joisel F, Raoult JP, et al: Serum concen-

ration of human alpha 2-HS glycoprotein during the inflam-atory process: Evidence that alpha 2-HS glycoprotein is a

egative acute-phase reactant. J Clin Invest 64:1118-1129, 197996. Daveau M, Davrinche C, Julen N, et al: The synthesis of

lpha-2-HS glycoprotein is down-regulated by cytokines inepatoma HepG2 cells. FEBS Lett 241:191-194, 198897. Avbersek-Luznik I, Malesic I, Rus I, et al: Increased

evels of osteoprotegerin in hemodialysis patients. Clin Chem0:1019-1023, 200298. Kazama JJ, Shigematsu T, Yano K, et al: Increased

irculating levels of osteoclastogenesis inhibitory factor (osteo-rotegerin) in patients with chronic renal failure. Am J Kidneyis 39:525-532, 200299. Haas M, Leko-Mohr Z, Roschger P, et al: Osteoprotegrin

nd parathyroid hormone as markers of high-turnover osteodys-rophy and decreased bone mineralization in hemodialysis pa-ients. Am J Kidney Dis 39:580-586, 2002

100. Coen G, Ballanti P, Balducci A, et al: Serum osteopro-egerin and renal osteodystrophy. Nephrol Dial Transplant 17:33-238, 2002