cartilage-derived growth factor and antitumor factor: past, present, and future studies

B

CP

FO

R

eorcttrmpbbtcfctuTfdmnmstctte(

aci(hih

Biochemical and Biophysical Research Communications 259, 1–7 (1999)

Article ID bbrc.1999.0610, available online at http://www.idealibrary.com on

REAKTHROUGHS AND VIEWS

artilage-Derived Growth Factor and Antitumor Factor:ast, Present, and Future Studies

ujio Suzukisaka University, 1-1, Yamadaoka, Suita, Osaka 565-0871, Japan

eceived March 30, 1999

The process replacing cartilage by bone is known as bovine cartilage inhibited the growth of solid tumorsaccTaTdaflCat

C

odcmgicFclp

ctsnceaac

ndochondral bone formation. During the initial stagef this process, epiphyseal cartilage cells proliferateapidly, becoming enlarged and hypertrophic, and in-rease matrix synthesis, such as cartilage-specific pro-eoglycans and type II, IX, X, and XI collagens. Duringhis series of events, the metabolism of cartilage isegulated by various hormones and locally generatedodulators. In 1972, Nevo and Dorfman (1) first re-

orted that exogenous proteoglycans extracted fromovine cartilage stimulate proteoglycan biosynthesisy cultured chondrocytes. Since then many investiga-ors have tried to identify other regulatory factors inartilage. Klagsbrun et al. (2) purified a potent growthactor from neonatal bovine cartilage and named itartilage-derived growth factor (CDGF). However,heir CDGF was later found to be identical with thebiquitous fibroblast growth factor-2, FGF-2 (3).herefore, cartilage contains intrinsic FGF-2 and this

actor has been suggested to be important during chon-rocyte growth. In 1980-81, we found some somato-edin-like factors in fetal bovine cartilage, which weamed cartilage-derived factor (CDF) (4-6). In theeantime, TGF-b and other members of its family,

uch as BMP (7), CDMP-1 and CDMP-2 (8), were iden-ified in cartilage and shown to be active in chondro-yte differentiation. Then, we succeeded in identifyinghe structures of unique cartilage-generated modula-ors of both chondrogenesis and subsequent osteogen-sis, and named these chondromodulin-I (ChM-I), -IIChM-II), and -III (ChM-III) (9-11).

On the other hand, cartilage is normally avascularnd resistant to invasion by both vascular endothelialells and neoplastic cells. Because of these character-stics, Kuettner’s group (12-14) and Folkman’s group15,16) tried to purify factors from cartilage that in-ibit proliferation of endothelial cells in vitro or tumor-

nduced angiogenesis in vivo. We also reported that aigh molecular weight fraction extracted from fetal

1

nd tumor-induced angiogenesis in the chick embryohorioallantoic membrane: we named this factorartilage-derived anti-tumor factor (CATF) (17,18).hese findings show that CATF has anti-angiogenicctivity, thereby inhibiting the growth of solid tumors.hen, the Harvard group partially purified a cartilage-erived collagenase inhibitor (CDI) from cartilage (19)nd a cartilage-derived angiogenesis inhibitor (ChDI)rom the conditioned medium of bovine scapular carti-age (20). However, we re-examined purification of oldATF preparations and found that the N-terminalmino acid sequence of the active anti-angiogenic frac-ion agrees with that of the ChM-I sequence (11).

ARTILAGE-DERIVED GROWTH FACTOR

Table I presents a chronological scheme of the devel-pment of our understanding of the details of cartilage-erived growth factors. Chondrocytes are highly spe-ialized for the syntheses and secretions of cartilageatrix, chondromucoprotein (proteoglycan) and colla-

en. In 1975, we showed that growth-cartilage cellssolated from the ribs of rats formed differentiatedolonies producing cartilage-like tissue in vitro (21).urthermore, the synthesis of sulfated glycosaminogly-ans by growth-cartilage cells was markedly stimu-ated by parathyroid hormone, calcitonin, and multi-lication-stimulating activity (22,23).

Functional matrix. Besides these hormones, extra-ellular matrix components have long been suggestedo play important roles in the control of proteoglycanynthesis. Nevo and Dorfman (1) showed that exoge-ous chondromucoprotein extracted from bovine nasalartilage stimulates proteoglycan synthesis by chickmbryo chondrocytes in suspension culture. Kosher etl. (24) demonstrated that extracellular proteoglycanggregates, obtained from embryonic chick cartilage,an stimulate chondrogenic expression by chick somite

0006-291X/99 $30.00Copyright © 1999 by Academic PressAll rights of reproduction in any form reserved.

etcnfttiipvpsictc

cbDciwTfcaacts

rsDtdtfwwcwwEiatDeldtu

FmctpcthM

TABLE I

1

1

1

111111

1

1

1

Vol. 259, No. 1, 1999 BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

xplants. In addition, Solursh and Meier (25) foundhat medium conditioned by chick embryo chondro-ytes stimulates the syntheses of sulfated glycosami-oglycans and collagen by the same cell type. There-ore, factors produced by chondrocytes or reserved inhe cartilage matrix may regulate the synthesis ofheir own extracellular matrix by a mechanism of pos-tive feedback. Then, we tried to isolate and character-ze these active factors from fetal bovine cartilage. Arotease-sensitive factor was extracted from fetal bo-ine cartilage with 1 M guanidine-hydrochloride andartially purified. This cartilage-derived factor (CDF)timulated proteoglycan synthesis and DNA synthesisn a dose-dependent manner in rat and rabbit costalhondrocytes in culture (26). These findings suggesthat CDF is involved in expression of the differentiatedhondrocytes.

FGF. Klagsbrun et al. (2, 26) partially purified aartilage-derived growth factor (CDGF) from neonatalovine scapula cartilage and found that it stimulatesNA synthesis and cell division of bovine chondro-

ytes. However, their CDGF was later found to bedentical to the ubiquitous growth factor, FGF-2 (3),hich is known as the most potent angiogenic factor.herefore, cartilage contains intrinsic FGF-2 and this

actor has been suggested to be important for chondro-yte growth. In contrast, we found that 11 kDa CDFnd 16 kDa CDF increase not only sulfation of glycos-minoglycans but also proliferation of rabbit chondro-ytes in culture (4-6, 27, 28). These findings indicatehat fetal bovine cartilage contains factors that showomatomedin-like activity in monolayer cultures of

Chronological Times of Results o

972 Chondromucoproteins extracted from bovine nasal cchondrocytes.

973 Extracellular proteoglycan aggregates from embryoby chick somite explants.

973 Medium conditioned by exposure to chick embryo chglycosaminoglycans by the same cell type.

977 Neonatal bovine cartilage contains a potent growth980 Purified CDGF stimulates growth of chondrocytes.980 Fetal bovine cartilage contains somatomedin-like gr983, 1986 CDF has synergistic effects with FGF in stimulatin986 CDGF is identical to the ubiquitous FGF-2.991 A new class of cartilage-specific matrix, chondromod

chondrocytes in the presence of FGF.996 Another growth-promoting component, chondromod

Chondromodulin-III (ChM-III) was purified from caprotein L31.

997 ChM-I stimulates colony formation of chondrocytesChM-I and ChM-II stimulate osteoblast proliferatioChM-I is localized in the avascular zone of epiphyse

endothelial cells as well as tube formation.998 FGF-2, TGF-b, and PTHrP down-regulate the expre

2

abbit chondrocytes. Then, we found that CDF hasynergistic effects with FGF or EGF in stimulatingNA synthesis in chondrocytes (29-31). The mode of

he combined action of CDF with FGF or EGF in chon-rocytes was studied by sequential treatments withhese factors. Both FGF and EGF had synergistic ef-ects with CDF in stimulating DNA synthesis, evenhen added 10 hr after exposure to CDF. Synergismas also observed in cells pretreated with CDF. These

ultures were treated for 5 hr with CDF and thenashed and treated with FGF or EGF (31). However,hen CDF was added more than 5 hr after FGF orGF, no synergism of effects was observed. These find-

ngs suggest that the cultured chondrocytes becamectivated to respond synergistically with the ubiqui-ous growth factors FGF and EGF for commitment toNA synthesis when they were exposed to CDF at anarly stage. Therefore, the designation “chondromodu-in” was proposed to supplant the older term “cartilage-erived factor (CDF)”. This term reflects the modula-ion of chondrocytes to respond synergistically to thesebiquitous growth factors.

Chondromodulin-I (ChM-I). As described earlier,GF-2 in combination with some unique growth-odulating factors in cartilage may act on chondro-

ytes. We, therefore, searched for such factors in car-ilage. In the course of the initial screening of growth-romoting components in extracts of fetal bovineartilage, we found at least two distinct factors inerms of their affinity for heparin: one eluted from aeparin-Sepharose column with buffer containing 0.5

NaCl and the other eluted with buffer containing 1.2

artilage-Derived Growth Factor

References

ilage stimulate proteoglycan synthesis by cultured 1

chick cartilage stimulate chondrogenic expression 24

drocytes stimulates the synthesis of sulfated 25

tor (cartilage-derived growth factor: CDGF). 226

th factors (cartilage-derived factor: CDF). 4-6NA synthesis in cultured chondrocytes 29-31

3n-I (ChM-I) stimulates growth of cultured 9

-II (ChM-II), is homologous to MIM-1. 10age and found to be identical with ribosomal 10

garose culture. 3236

cartilage and inhibits proliferation of vascular 11, 42

on of ChM-I mRNA in chondrocytes. 33

n C

art

nic

on

fac

owg D

uli

ulinrtil

in an.al

ssi

M NaCl. Finally we purified these active componentstcntcstwuc“tpCclcpad

spptNappsnimcNsiprcaTNtao“peCtmgaps

A single ChM-I mRNA of 1.7 kb was identified inptvC

tmtmlpozdiccgcdCcn

vaTlptFi

pstaNcqecocmpHtcdcpdsm


hat function in growth modulation of differentiatedhondrocytes (9). The first component identified was aovel 25 kDa glycoprotein expressed specifically in car-ilage that stimulated the proliferation of chondro-ytes. In the presence of FGF-2 (0.4 ng/ml), this factorynergistically stimulated DNA synthesis 30-fold overhe control. Its activity was detectable at 1 ng/ml andas maximal at about 200 ng/ml. Furthermore, it stim-lated the synthesis of cartilage-proteoglycans byhondrocytes. We renamed this unique principlechondromodulin-I (ChM-I) (9), because it modulatedhe growth of cultured chondrocytes, and cartilage wasresumed to contain some additional modulators.hM-I also stimulated colony formation of chondro-

ytes cultured in soft agar (32). ChM-I alone stimu-ated the formation of chondrocyte colonies weakly, butaused marked stimulation of colony formation in theresence of FGF-2. Therefore, ChM-I participates in anutocrine signaling mechanism for anchorage-indepen-ent growth of chondrocytes in vitro (32).Next, we cloned cDNA of ChM-I. Mature ChM-I con-

ists of 121 amino acids and is coded as the C-terminalart of a larger precursor which has 335 amino acids. Arocessing signal, RERR, is located just before the ma-ure ChM-I sequence. Glutamic acid is the only-terminal residue. Two O-glycosylation sites (Thr9

nd Thr22) and one N-glycosylation site (Asn30) areresent in the mature ChM-I sequence. The ChM-Irecursor contains a single hydrophobic region and atretch of 27 amino acid residues from Gly43 to Ile59

ear its N-terminal. This region is presumed to benvolved in membrane insertion. We speculate that

ature ChM-I is secreted after processing of the pre-ursor at the preceding processing signal and that the-terminal two-thirds of the molecule reside on the cell

urface of chondrocytes. Sequence similarity analysisndicated that the N-terminal two-thirds of the ChM-Irecursor have significant similarity with human andat pulmonary surfactant apoprotein C, which is spe-ifically expressed in alveolar type II cells and contains

transmembrane domain in the N-terminal region.his sequence similarity leads us to speculate that the-terminal part of the ChM-I precursor may also con-

ribute to functional regulations of cartilage cells, suchs possible mediation of the growth-promoting actionf ChM-I. We termed this putative membrane proteinchondrosurfactant protein (Ch-SP)”. Then, we ex-ressed the ChM-I in COS cells and found that thexpressed product recovered from the supernatant ofOS cells gives a similar diffuse band corresponding to

hat of authentic ChM-I on western blot analysis. Thiseans that the expressed product is secreted after

lycosylation in a similar manner to authentic ChM-Ind then processing of the precursor protein at thereceding processing signal. Ch-SP remains on the cellurface.

3

olyadenylated bovine epiphyseal cartilage RNA. In-erestingly, northern blot analysis of total RNA fromarious bovine tissues indicated that expression ofhM-I mRNA is highly specific to cartilage.Next, we examined the immunolocalization of ma-

ure ChM-I and the in situ hybridization of ChM-IRNA using caudal vertebral growth-plates of bovine

ail. Results indicated that ChM-I protein and itsRNA are localized only in avascular zones of carti-

age; i.e. the resting, proliferating, and early hypertro-hic zones, and are not detectable in the outer surfacef cartilage or in the late hypertrophic and calcifyingones which allow vascular invasion during endochon-ral bone formation (11). Immunohistochemical stain-ng also showed that ChM-I is secreted from chondro-ytes and accumulated in the inter-territorial matrix ofartilage (11). In addition, no transcript of the ChM-Iene was detected in undifferentiated ATDC5 mousehondrogenic cells, but ChM-I transcripts were readilyetectable in differentiated cultures of ATDC5 cells.hM-I mRNA was also expressed in cultured rabbithondrocytes, and resting cartilage of rabbit ribs, butot in osteoblastic MC3T3-E1 cells.As described before, cartilage produces FGF-2, a

ery potent angiogenic factor, although cartilage is anvascular tissue that is rarely invaded by neoplasms.o study puzzling problem, we examined the immuno-

ocalization of FGF-2 and confirmed that it was ex-ressed not only in cartilage, but also in other connec-ive tissues. But in contrast to ChM-I, the staining ofGF-2 was stronger in the calcified cartilage zone and

n bone. Thus, ChM-I may counteract FGF-2 function.Shukunami and Hiraki (33) isolated rabbit ChM-I

recursor cDNA by RT-PCR. The deduced amino acidequence of rabbit mature ChM-I was compared withhat of its human counterpart. One amino acid deletionnd 12 amino acid substitutions were found near the-glycosylation site (Asn29), although the glycosylation

onsensus sequence was still valid in the rabbit se-uence (33). Northern blot analysis revealed that thexpression of ChM-I mRNA occurred specifically inartilage. Moreover, BMP-2 stimulated the expressionf the differentiated phenotype of primary culturedhondrocytes (7) and up-regulated type II collagenRNA in chondrocytes (33). This result was well com-

atible with the chondrogenic activity of BMP-2 (34).owever, BMP-2 or activin only marginally affected

he level of ChM-I mRNA in chondrocytes (33). Inontrast, FGF-2, TGF-b2, and PTHrP each markedlyown-regulated the expression of ChM-I mRNA inhondrocytes (33). These results indicate that the ex-ression of ChM-I is controlled by local growth andifferentiation factors. A higher level of ChM-I expres-ion was maintained in mature chondrocytes, and wasarkedly down-regulated on shift of the metabolic

state of chondrocytes from maturing to proliferatingc

chiwPWCchSn5uTwcudTmakDrosictoectbcpt(CTcmtpc(pajbTCp

erated, osteopoietic coupling factors of cartilage-boneg

mcCaraglCfflfa

C

r1todTalpaKigcla

ctnpamtoXofgitWjw


ells by growth stimulation with FGF-2 and/or TGF-b.

Chondromodulin-II (ChM-II). We purified anotherartilage-derived modulator with a lower affinity toeparin (0.5% NaCl eluate) to homogeneity and named

t chodromodulin-II (ChM-II) (10). Final purificationas achieved by reversed phase HPLC. On SDS-AGE, the active fraction gave a single band of 16 kDa.e determined the complete amino acid sequence ofhM-II and found that it consists of 133 amino acids,

ontains one tryptophan residue near the N-terminus,as six cysteine residues, and is rich in lysine residues.creening of a data base indicated that ChM-II has aovel amino acid sequence. However, it has a sequence7% identical to the repeat 1 and repeat 2 of the prod-ct of the chicken promyelocyte-specific mim-1 gene.he positions of six half-cystine residues in ChM-IIere all conserved in these MIM-1 proteins. mim-1

DNA was first reported by Ness et al. (35). The prod-ct of the chicken promyelocyte-specific mim-1 gene isirectly transactivated by the v-Myb oncogene product.he MIM-1 protein is stored in the a-granules of nor-al immature granulocytes and expressed as 35 kDa

nd 16 kDa forms. However, its function is not yetnown. Like ChM-I, purified ChM-II stimulated bothNA synthesis and proteoglycan synthesis of cultured

abbit chondrocytes dose-dependently. In the presencef FGF-2, ChM-II also synergistically stimulated DNAynthesis of chondrocytes. However, ChM-II did notnhibit DNA synthesis of cultured vascular endothelialells of bovine carotid artery, whereas ChM-I did. In-erestingly, ChM-II stimulated the differentiation ofsteoclasts. The effects of ChM-II on osteoclast differ-ntiation were evaluated using an unfractionated boneell culture system containing mature osteoclasts fromhe femur and tibia of newborn mice. When mouseone cells were cultured on dentin slices, ChM-II in-reased the total area and the number of resorptionits dose-dependently. Osteoclasts were represented ashe number of tartarate-resistant acid phosphataseTRAP)-positive multinucleated cells per unit area.hM-II also dose-dependently increased the number ofRAP-positive multinucleated cells in the same con-entration range. In these systems, an activated vita-in D3 was not necessary to induce osteoclast forma-

ion. Furthermore, ChM-I and ChM-II stimulated theroliferation of clonal mouse osteoblastic MC3T3-E1ells as well as primary mouse osteoblasts in culture36). These findings suggest that epiphyseal growthlate cartilage may be functionally involved in thective proliferation of osteoblasts at the osteochondralunction, and hence in the longitudinal growth of longones by sending a specific growth-promoting signal.herefore, the cartilage-derived matricrine factorshM-I and ChM-II are both chondormodulin/osteo-oietin, and are the first characterized naturally gen-

4

rowth in endochondral bone formation.

Chondromodulin-III (ChM-III). We also found ainor component with higher affinity for heparin,

hondromodulin-III (ChM-III), in cartilage extracts.hM-III stimulated DNA synthesis in chondrocytes,nd its N-terminal sequence was identical with that ofibosomal protein L31 but lacked the three N-terminalmino acids (10). These findings suggest that therowth and differentiation of chondrocytes are regu-ated by multiple components, such as FGF-2, ChM-I,hM-II, and ChM-III, in the cartilage matrix. There-

ore, the above findings indicate that these variousunctional components participate sequentially in cel-ular growth and differentiation in endochondral boneormation in autocrine and matricrine fashions andlso in punctually and spatially regulated manners.

ARTILAGE-DERIVED ANTITUMOR FACTOR

Table II shows the chronological order of progress inesearch on the cartilage-derived anti-tumor factor since973. Kuettner’s group (12) first demonstrated that car-ilage is avascular and resistant to invasion by neoplasmsr inflammatory processes. After extraction with guani-ine-hydrochloride, the tissue loses this resistance (13).hese facts have led some investigators to try to extractnd purify factors from cartilage that inhibit neovascu-arization and tumor invasion. Folkman’s group (15, 16)artially purified angiogenesis inhibitors from cartilagend showed their inhibitory activity on tumor growth.uettner’s group (14) then reported extractable protease

nhibitors from bovine cartilage which inhibited tumorrowth in mice. At that time, we were trying to purifyartilage-derived growth factors from fetal bovine carti-age, and used the rest of the material for study of anti-ngiogenesis factors.

Cartilage-derived anti-tumor factor (CATF). Fetalalves were homogenized and the homogenate was ex-racted with 1 M guanidine-hydrochloride. The super-atant was fractionated with acetone. The materialrecipitated with 45-65% acetone was filtered throughn Amicon XM300 filter, which excludes molecules ofore than 3 3 105 daltons. The filtrate was concen-

rated on a Toyo UP20 filter, which excludes moleculesf less than 2 3 104 daltons and the concentrate (UP20-M300) was dialyzed against distilled water, and ly-philized. First, we examined the effects of the acetoneraction (45-65%) and UP20-XM300 fraction on therowth of solid sarcoma 180 inoculated subcutaneouslynto the loins of ICR/CRJ mice. In control mice, theumors grew slowly in the first week, and then rapidly.

hen 2 mg of the acetone fraction (45-65%) was in-ected twice in the first week, the growth of the tumoras not inhibited. However, when 2 mg of the UP20-

Xt5wpwdmcttFit

chiiaD(pTBtmrtashd

d(hcpdioTdaeariaqdc

lotgriaicwat

TABLE II

11111

1

1

1

11


M300 fraction was injected twice in the first week,umor growth was inhibited 60.7% and 56% after 3 and

weeks, respectively (17). The effect of this fractionas striking, because solid tumors shrank and stoppedroliferating in 43% of the treated mice after 3 and 5eeks. Therefore, we named this fraction cartilage-erived anti-tumor factor (CATF). In contrast, whenice were given an intraperitoneal inoculation of sar-

oma 180 cells followed by intraperitoneal injection ofhe same doses of CATF, no difference was observed inhe survival times of treated and control animals (17).ifty percent of the mice died 17 days after tumor

noculation in both the control group and the CATF-reated group.

When CATF was added to sarcoma 180 cultures at aoncentration of 200 mg/ml, it did not significantly in-ibit the proliferation of the cells. This indicates that

nhibition by CATF of the growth of solid sarcoma 180s not due to its direct effect on sarcoma 180 cells. Welso showed that CATF inhibits the proliferation andNA synthesis of bovine pulmonary artery endothelial

BPAE) cells in culture (37). Next, CATF was furtherurified by DEAE-Sepharose CL-6B chromatography.he main peak with activity on DNA synthesis inPAE cells inhibited the growth of B16 melanoma

ransplanted into the paws of C57BL/6N mice and B16elanoma-induced angiogenesis in chick embryo cho-

ioallantoic membranes (18). These findings suggesthat CATF is an anionic macromolecule and has anti-ngiogenic activity, thereby inhibiting the growth ofolid tumors. Furthermore, we showed that a clonaluman chondrosarcoma cell line (HCS-2/8) (38) pro-uces an anti-angiogenic antitumor factor (39).

Chronological Times of Results on

973 Cartilage is avascular and resistant to invasion by ne975 After extraction with guanidine-HCl, cartilage loses t976–1980 Angiogenesis inhibitors extracted from cartilage inhib977 Protease inhibitors extracted from cartilage inhibit tu984 High molecular weight fractions extracted from fetal

(cartilage-derived anti-tumor factor: CATF).988 Partially purified CATF inhibits the proliferation of e

B16-melanoma-induced angiogenesis.990 A clonal human chondrosarcoma cell line produces an

Cartilage-derived collagenase inhibitor (CDI) is an an992 The N-terminal amino acid sequence of cartilage-deri

conditioned medium of bovine scapular cartilage ismetalloproteinase-1 (TIMP-1).

995 The N-terminal amino acid sequence of the anti-angio997 The N-terminal amino acid sequence of endothelial ce

indicates that the inhibitor is identical with chondrChM-I is localized in the avascular zone of epiphysealRecombinant ChM-I stimulates DNA synthesis and p

proliferation of vascular endothelial cells as well as

5

Cartilage-derived inhibitor (CDI) and chondrocyte-erived angiogenesis inhibitor (ChDI). Moses et al.

19) demonstrated that cartilage contains a potent in-ibitor of neovascularization in vivo by chick embryohorioallantoic membrane assay and rabbit cornealocket assay. They named this inhibitor cartilage-erived inhibitor (CDI). The sequence of 28 amino ac-ds in the N-terminal part of CDI is identical with thatf the tissue inhibitor of metalloproteinase-1 (TIMP-1).hen, Moses et al. (20) reported that conditioned me-ium of bovine scapular chondrocytes produces annti-angiogenic factor (chondrocyte-derived angiogen-sis inhibitor: ChDI) which shows the same biologicalctivities as CDI. On the other hand, we previouslyeported that TIMP-1 strongly inhibited angiogenesisn chick yolk sac membranes induced by spermine (40)nd showed that the N-terminal 11 amino acid se-uence of the angiogenesis inhibitor from culture me-ium conditioned by a human chondrosarcoma-derivedell line, HCS-2/8, is identical with that of TIMP-2 (41).

ChM-I makes a comeback at last. However, in theate 1990s, a major breakthrough occurred in the fieldf cartilage-derived growth factors and anti-tumor fac-ors. As reported earlier (17, 18), CATF inhibited an-iogenesis and the growth of solid tumors. We thene-examined the anti-endothelial cell growth activityn CATF preparations and found that the inhibitoryctivity is adsorbed to a heparin column and recoveredn the 0.5-1.2 M NaCl eluate. When this fraction washaracterized by gel filtration, its apparent moleculareight was decreased to 10-30 kDa. Interestingly,mino acid sequence analysis of the N-terminal part ofhe active fraction showed that it is identical to that of

rtilage-Derived Antitumor Factor

References

asms. 12resistance. 13umor growth. 15, 16r growth. 14ine cartilage inhibit the growth of solid sarcoma-180 17

thelial cells, the growth of B16 melanoma and 18

ti-angiogenic antitumor factor. 39genesis inhibitor. 19angiogenesis inhibitor (ChDI) purified from the

ntical with that of the tissue inhibitor of20

ic antitumor factor is identical with that of TIMP-2. 41rowth inhibitor extracted from fetal bovine cartilageodulin-I (ChM-I).

11

rtilage. 11oglycan synthesis of chondrocytes, but inhibitse formation.

11, 42

Ca

oplhisit tmobov

ndo

angiovedide

genll gomca

rotetub

the ChM-I sequence (11). We then found that purifiedalmc(

C

csoboihowaSimBcafiivcmiolccoCwfilt

tatafpto

CCsri

thelial cells. ChM-I counters the effect of FGF-2 andtucmsfpwsmigiivsgeslnwwC

tdvpfitC

qstutqlceCtvd

A

AfY


nd recombinant ChM-I itself inhibited both the pro-iferation of vascular endothelial cells and vessel for-

ation by bovine endothelial cells cultured betweenollagen gels by a three-dimensional sandwich method11, 42).

ONCLUDING REMARKS AND PERSPECTIVES

During endochondral bone formation, epiphysealartilage cells proliferate and increase matrix synthe-is, becoming enlarged and hypertrophic. Calcificationf cartilage and replacement by bone are accompaniedy blood vessel invasion. To characterize the functionf ChM-I in vivo, we examined its effect on BMP-nduced ectopic bone formation. BMP was adsorbed oneparin-Sepharose beads and inoculated subcutane-usly into nude mice. Endochondral bone formationith extensive vascularization occurred three weeksfter BMP administration. In contrast, when heparin-epharose beads with adsorbed BMP and ChM-I were

njected, vascularization and endochondral bone for-ation were markedly inhibited and extensive Alcianlue-positive cartilage remained. These findings indi-

ate that ChM-I may be named chondromodulin/ngioinhibin. Thus ChM-I was the first clearly identi-ed natural modulator and anti-angiogenic factor

solated from cartilage. ChM-I inhibits the invasion ofascular endothelial cells, thereby further stimulatingartilage growth and inhibiting premature replace-ent of cartilage by bone in the fetal stage. This is

mportant in preventing dwarfism. These findings mayffer a clue to the longstanding question of why carti-age is an avascular tissue, in spite of its abundantontent of the potent angiogenic factor FGF-2. Re-ently, we cloned human ChM-I. Sequence similaritiesf the entire precursor molecule, Ch-SP, and maturehM-I of bovine ChM-I with those of human ChM-Iere 92.2%, 95.8%, and 80.0%, respectively. We con-rmed that recombinant human ChM-I also stimu-

ated the proliferation of chondrocytes, but inhibitedhe proliferation of vascular endothelial cells.

Solid tumors grow slowly in avascular conditions:hey do not generally exceed a few millimeters in di-meter because of limited supplies of oxygen and nu-ritients. But when new capillaries develop in the hostnd reach the tumors, the tumors grow rapidly. There-ore, anti-angiogenesis might be a rational approach inreventing solid tumor growth. ChM-I, recently iden-ified as a cartilage-generated anti-angiogenic factor isne of the most promising factors for this purpose.In the Open Forum of the US-Japan Cooperativeancer Research Program at the Natcher Conferenceenter on the Bethesda NIH Campus in 1996, I pre-ented results on ChM-I research in a session on theole of cytokines in modulating angiogenesis. ChM-Inhibits the proliferation and tube formation by endo-

6

he bone forming activity of BMP. Thus, ChM-I stim-lates cartilage formation and prevents its prematureonversion to bone. The angiostatic properties of ChM-Iay make it a nontoxic antitumor agent. In the same

ession, O’Reilly presented results of studies per-ormed in Folkman’s laboratory. He has found thatrimary tumors generate an inhibitor of angiogenesishich suppresses the growth of tumor metastases. A

uppressive factor present in plasma and urine of tu-or bearing mice was purified and identified as an

nternal fragment of plasminogen, and named angio-enesis inhibitor angiostatin (43). When given system-cally, angiostatin strongly inhibited the angiogenesisnduced by FGF-2 in a mouse model of corneal neo-ascularization. Systemic therapy with angiostatintrongly inhibited both metastatic and primary tumorrowth of murine and human tumors in mice. How-ver, plasminogen itself was inactive. Another angio-tatic factor purified from a murine hemangioendothe-ioma is a fragment of type XVIII collagen and has beenamed endostatin (44). Endostatin, in combinationith angiostatin, eradicated various tumors in miceithout inducing resistance or obvious side effects.artilage may contain these types of factors.Psoriasis is an inflammatory skin disease with dila-

ion of capillaries as an early histological change. Moreeveloped psoriatic lesions show proliferation of bloodessels and neovascularization. Dupont et al. (45) re-orted that the anti-angiogenic agent AE-941, derivedrom shark cartilage, has anti-angiogenic and anti-nflammatory properties, and thus it is promising forreatment of psoriasis. Shark cartilage may contain ahM-I-like factor.Recently, Okihana and Yamada (46) prepared a good

uality cDNA library from mouse growth cartilage,elected 1401 clones from this library, and examinedheir sequences. Among 608 clones examined, 196 werenknown and 2 were only poly A. The remaining two-hirds of the clones (410/608) coded for known se-uences. Of these, 55 clones coded for type II (pro)col-agen, 54 for osteonectin, and 22 for other cartilageollagens (type IX, type X, and type XI). The othersncode cartilage extracellular matrix genes such ashM-I and general cellular genes. Almost one-third of

he genes were novel. Their cDNA libraries will bealuable for study of other unidentified cartilage-erived growth factors and anti-tumor factors.

CKNOWLEDGMENTS

I thank my colleagues Y. Hiraki, H. Inoue, C. Shukunami, H.,kiyama, J. Kondo, K. Iyama, and M. Kumegawa for untiring ef-

orts. I also thank my former colleagues, Y. Kato, M. Takigawa, and. Yugari, from whom I have learned so much.

REFERENCES

1

1

1

1

1

1

1

1

1

1

2

2

2

2

24. Kosher, R. A., Lash, J. W., and Minor, R. R. (1973) Dev. Biol. 35,

22

2

2

2

3

3

3

3

3

33

3

3

3

4

4

4

4

4

4

4


1. Nevo, Z., and Dorfman, A. (1972) Proc. Natl. Acad. Sci. USA 69,2069–2072.

2. Klagsbrun, M., Langer, R., Levenson, R., Smith, S., and Lillehei,C. (1977) Exp. Cell Res. 105, 99–108.

3. Lobb, R., Sasse, J., Shing, Y., D’Amore, P., Sullivan, R., Jacobs,J., and Klagsbrun, M. (1986) J. Biol. Chem. 261, 1924–1928.

4. Kato, Y., Nomura, Y., Daikuhara, Y., Nasu, N., Tsuji, M., Asada,A., and Suzuki, F. (1980) Exp. Cell Res. 130, 73–81.

5. Kato, Y., Nomura, Y., Tsuji, M., Ohmae, H., Kinoshita, M.,Hamamoto, S., and Suzuki, F. (1981) Exp. Cell Res. 132, 339–347.

6. Kato, Y., Nomura, Y., Tsuji, M., Kinoshita, M., Ohmae, H., andSuzuki, F. (1981) Proc. Natl. Acad. Sci. USA 78, 6831–6835.

7. Hiraki, Y., Inoue, H., Shigeno, C., Sanma, Y., Bentz, H., Rosen,D. M., Asada, A., and Suzuki, F. (1991) J. Bone Miner. Res. 6,1373–1385.

8. Chang, S. S., Hoang, B., Thomas, J. T., Vukicevic, S., Luyten,F. P., Ryba, N. J., Kozak, C. A., Reddi, A. H., and Moos, M. Jr.(1994) J. Biol. Chem. 269, 28227–28234.

9. Hiraki, Y., Tanaka, H., Inoue, H., Kondo, J., Kamizono, A., andSuzuki, F. (1991) Biochem. Biophys. Res. Commun. 175, 971–977.

0. Hiraki, Y., Inoue, H., Kondo, J., Kamizono, A., Yoshitake, Y.,Shukunami, C., and Suzuki, F. (1996) J. Biol. Chem. 271, 22657–22662.

1. Hiraki, Y., Inoue, H., Iyama, K., Kamizono, A., Ochiai, M.,Shukunami, C., Iijima, S., Suzuki, F., and Kondo, J. (1997)J. Biol. Chem. 272, 32419–32426.

2. Eisenstein, R., Sorgente, N., Soble, L. W., Miller, A., and Kuett-ner, K. E. (1973) Am. J. Pathol. 73, 765–774.

3. Sorgente, N., Kuettner, K. E., Soble, L. W., and Eisenstein, R.(1975) Lab. Invest. 32, 217–212.

4. Kuettner, K. E., Soble, N., Croxen, R. L., Marczynska, B., Hiti,J., and Harper, E. (1977) Science 196, 653–654.

5. Langer, R., Brem, H., Falterman, K., Klein, M., and Folkman, J.(1976) Science 195, 70–72.

6. Langer, R., Coon, H., Vacanti, J., Haudenshild, C., and Folkman,J. (1980) Proc. Natl. Acad. Sci. USA 77, 4331–4335.

7. Suzuki, F., Takigawa, M., Hiraki, Y., Kato, Y., and Yugari, Y.(1984) J. Bone Miner. Metab. 2, 231–235.

8. Takigawa, M., Shirai, E., Enomoto, M., Hiraki, Y., Suzuki, F.,and Yugari, Y. (1988) J. Bone Miner. Metab. 6, 83–92.

9. Moses, M. A., Sudhalter, J., and Langer, R. (1990) Science 248,1408–1410.

0. Moses, M. A., Sudhalter, J., and Langer, R. (1992) J. Cell Biol.119, 475–482.

1. Shimomura, Y., Yoneda, T., and Suzuki, F. (1975) Calcif. TissueRes. 19, 179–187.

2. Suzuki, F., Yoneda, T., and Shimomura, Y. (1976) FEBS Lett. 70,155–158.

3. Kato, Y., Nasu, N., Takase, T., Daikuhara, Y., and Suzuki, F.(1980) Exp. Cell Res. 125, 167–174.

7

210–220.5. Solursh, M., and Meier, S. (1973) Dev. Biol. 30, 279–289.6. Klagsbrun, M., and Smith, S. (1980) J. Biol. Chem. 255, 10859–

10866.7. Suzuki, F., Hiraki, Y., and Kato, Y. (1987) Methods Enzymol.

146, 313–320.8. Kato, Y., Nakashima, K., Sato, K., Yan, W. A., Iwamoto, M.

(1991) Methods Enzymol. 198, 416–424.9. Kato, Y., Hiraki, Y., Inoue, H., Kinoshita, M., Yutani, Y., and

Suzuki, F. (1983) Eur. J. Biochem. 129, 685–690.0. Hiraki, Y., Kato, Y., Inoue, H., and Suzuki, F. (1986) Eur. J. Bio-

chem. 158, 333–337.1. Hiraki, Y., Inoue, H., Kato, Y., Fukuya, M., and Suzuki, F. (1987)

Mol. Cell. Biochem. 76, 185–193.2. Inoue, H., Kondo, J., Koike, T., Shukunami, C., and Hiraki, Y.

(1997) Biochem. Biophys. Res. Commun. 241, 395–400.3. Shukunami, C., and Hiraki, Y. (1998) Biochem. Biophys. Res.

Commun. 249, 885–890.4. Shukunami, C., Ohta, Y., Sakuda, M., and Hiraki, Y. (1998) Exp.

Cell Res. 241, 1–11.5. Ness, S. A., Marknell, A., and Graf, T. (1989) Cell 59, 1115–1126.6. Mori, Y., Hiraki, Y., Shukunami, C., Kakudo, S., Shiokawa, M.,

Kagoshima, M., Mano, H., Hakeda, Y., Kurokawa, T., Suzuki, F.,and Kumegawa, M. (1997) FEBS Lett. 406, 310–314.

7. Takigawa, M., Shirai, E., Enomoto, M., Hiraki, Y., Fukuya, M.,Suzuki, F., Shiio, T., and Yugari, Y. (1985) Cell Biol. Int. Rep. 9,619–625.

8. Takigawa, M., Tajima, K., Pan, H. O., Enomoto, M., Kinoshita,A., Suzuki, F., Takano, Y., and Mori, Y. (1989) Cancer Res. 49,3996–4002.

9. Takigawa, M., Pan, H. O., Enomoto, M., Kinoshita, A., Nishida,Y., Suzuki, F., and Tajima, K. (1990) Anticancer Res. 10, 311–316.

0. Takigawa, M., Nishida, Y., Suzuki, F., Kishi, J., Yamashita, K.,and Hayakawa, T. (1990) Biochem. Biophys. Res. Commun. 171,1264–1271.

1. Ohba, Y., Goto, Y., Kimura, Y., Suzuki, F., Hisa, T., Takahashi,K., and Takigawa, M. (1995) Biochim. Biophys. Acta 1245, 1–8.

2. Hiraki, Y., Kono, T., Sato, M., Shukunami, C., and Kondo, J.(1997) FEBS Lett. 415, 321–324.

3. O’Reilly, M. S., Homgren, L., Chen, C., and Folkman, J. (1996)Nature Med. 2, 689–692.

4. O’Reilly, M. S., Boehm, T., Shing, Y., Fukui, N., Vaslos, G., Lane,W. S., Flynn, E., Birkhead, J. R., Olsen, B. R., and Folkman, J.(1997) Cell 88, 277–285.

5. Dupont, E., Savard, P. E., Jourdain, C., Juneau, C., Thibodeau,A., Ross, N., Marenus, K., Maes, D. H., Pelletier, G., and Sander,D. N. (1998) J. Cutan. Med. Surg. 2, 146–152.

6. Okihana, H., and Yamada, K. (1999) J. Bone Miner. Res. 14,304–310.

cartilage-derived growth factor and antitumor factor: past, present, and future studies

Documents