review nov (nephroblastoma overexpressed) and the ccn ... · (nephroblastoma overexpressed gene)....

Review

NOV (nephroblastoma overexpressed) and theCCN family of genes: structural and functionalissues

B Perbal

AbstractThe CCN family of genes presently con-sists of six distinct members encodingproteins that participate in fundamentalbiological processes such as cell prolifera-tion, attachment, migration, diVerentia-tion, wound healing, angiogenesis, andseveral pathologies including fibrosis andtumorigenesis. Whereas CYR61 andCTGF were reported to act as positiveregulators of cell growth, NOV (nephrob-lastoma overexpressed) provided the firstexample of a CCN protein with negativeregulatory properties and the first exam-ple of aberrant expression being associ-ated with tumour development. Thesubsequent discovery of the ELM1,rCOP1, and WISP proteins has broadenedthe variety of functions attributed to theCCN proteins and has extended previousobservations to other biological systems.This review discusses fundamental ques-tions regarding the regulation of CCNgene expression in normal and pathologi-cal conditions, and the structural basis fortheir specific biological activity. After dis-cussing the role of nov and other CCNproteins in the development of a variety ofdiVerent tissues such as kidney, nervoussystem, muscle, cartilage, and bone, thealtered expression of the CCN proteins invarious pathologies is discussed, with anemphasis on the altered expression of novin many diVerent tumour types such asWilms’s tumour, renal cell carcinomas,prostate carcinomas, osteosarcomas,chondrosarcomas, adrenocortical carci-nomas, and neuroblastomas. The possibleuse of nov as a tool for molecular medicineis also discussed. The variety of biologicalfunctions attributed to the CCN proteinshas led to the proposal of a model in whichphysical interactions between the aminoand carboxy portions of the CCN proteinsmodulate their biological activity andensure a proper balance of positive andnegative signals through interactions withother partners. In this model, disruptionof the secondary structure of the CCNproteins induced by deletions of either

terminus is expected to confer on thetruncated polypeptide constitutive posi-tive or negative activities.(J Clin Pathol: Mol Pathol 2001;54:57–79)

Keywords: cancer; diVerentiation; signalling; develop-ment; angiogenesis; fibrosis; ctgf; cyr61; wisp; CCN

The regulation of cellular growth, diVerentia-tion, and death is central to a myriad ofbiological processes that govern the appear-ance, maintenance, and termination of life.

Although many genes whose products arereported to play a key role in cell growth con-trol have been described, several basic regula-tory circuits involve proteins yet to be identi-fied.

The discovery of three structurally relatedgenes whose expression was modulated bygrowth factors and altered in cancer cells hasled to the recognition of a new family of cellgrowth regulators with unique biological prop-erties.

Bork1 first proposed that these genes shouldbe designated the CCN family of genes, whichstands for Cyr61 (cystein rich protein), Ctgf(connective tissue growth factor), and Nov(nephroblastoma overexpressed gene). Theexistence of this family of genes has recentlybeen reinforced by the discovery of three newmembers (see below).

For many years, the CCN proteins seemednot to share much more than their strikinglyconserved primary structure and no physi-ological or biological properties had beenclearly assigned to them.

The bulk of the recent results presented atthe first international workshop on the CCNfamily of genes and the increasing pace atwhich papers are being published in this fieldpoint to the increased attention being focusedon these proteins. Their roles are being clearlyestablished in fundamental biological proc-esses such as cell proliferation, attachment,migration, and diVerentiation, wound healing,angiogenesis, and several pathologies includingfibrosis and tumorigenesis. Furthermore, itappears that in spite of their highly conservedstructure, these proteins display non-redundant functions, which are apparently

J Clin Pathol: Mol Pathol 2001;54:57–79 57

Laboratoire d’Oncologie Virale etMoléculaire, UFR deBiochimie, UniversitéParis 7-D. Diderot, 2Place Jussieu, 75 005Paris, FranceB Perbal

Correspondence to:Professor [email protected]

Accepted for publication9 January 2001

www.molpath.com

on Decem

ber 6, 2020 by guest. Protected by copyright.

http://mp.bm

j.com/

Mol P

ath: first published as 10.1136/mp.54.2.57 on 1 A

pril 2001. Dow

nloaded from

http://mp.bmj.com/

required at diVerent locations and at diVerenttimes along the same biological pathways.

Understanding the structural and functionalbases for the molecular diversity of the CCNfamily is an important challenge. It will set thestage for understanding the role of these genesin the control of cellular proliferation, diVeren-tiation, and death, thereby allowing the use ofthese genes and proteins in modern biomedi-cal technologies for early diagnosis andtreatment.

Because recent publications have dealt indetail with the description of the CCN genesand corresponding proteins,2–4 this review willdeal with some fundamental questions regard-ing the expression of CCN proteins in normaland pathological conditions, and the structuralbasis for their specific biological activity, withan emphasis on studies performed on the novgene in my laboratory.

Isolation of the CCN genesThe CCN family of genes presently consists ofsix distinct members in the human (fig 1), theorthologues of which have been described inseveral diVerent species ranging from xenopusto rodents.

The way in which the CCN genes were iso-lated might provide valuable information forunderstanding their diVerential expression andbiological properties. The diVerent strategiesthat have led to the isolation of CCN genesinclude: subtractive hybridisation and diVeren-tial display of RNAs whose expression isaVected in normal cells stimulated to grow byserum, growth factors, and viruses, or intumour cells as compared with their normalcounterparts.

The diVerential screening strategies arebased on the fact that RNA species are underrepresented or lacking in one of the two situa-tions analysed. Although most authors at-tribute increased RNA values to stimulation oftranscription, it is important to keep in mindthat RNA metabolism can be aVected dra-matically by the events that lead to transfor-mation or induction of proliferation in re-sponse to viral infection and growth factors.Only in a few instances was it possible to checkwhether increased or decreased gene expres-sion was indeed responsible for the diVerentiallevels of transcripts detected, by means of

nuclear run on transcription assays. Thisaspect must be approached with extreme cau-tion because it has a pronounced impact onour way of thinking when analysing thebiological properties of potential regulatorymolecules, such as CCN proteins, whoseexpression is developmentally regulated in aspatiotemporal way.

GENES THAT WERE ISOLATED ON THE BASIS OF

THEIR DIFFERENTIAL EXPRESSION IN QUIESCENT

AND SERUM STIMULATED CELLS

Under starvation conditions (incubation in lowserum concentration), confluent cells stopgrowing and reach the G0 state of the cellcycle. G1 phase re-entry of the starved cellsupon treatment with serum triggers the stimu-lation of immediate early genes, whose expres-sion proceeds shortly after serum stimulation,without the need for de novo protein synthesis(the immediate early terminology was used bycomparison with the ordered expression ofviral genes upon infection of eukaryotic cells).Characterisation of these immediate genesprovided important clues as to the type ofgenes responsible for the regulation of masterswitches involved in the control of cell growth.

This experimental approach, which had pre-viously led to the discovery of a variety ofsignalling proteins and transcription factors,led to the isolation of the murine cyr61 andfisp12 genes.

Fibroblastic cells of diVerent origins havebeen used to analyse gene expression duringserum induced G0/G1 transition. DiVerentialscreening of a cDNA library prepared fromserum stimulated mouse BALB/c3T3 fibro-blasts led to the isolation of cyr61, whichencodes a secreted 379 residue polypeptide ofapparent molecular mass 42 kDa,5 whereas asimilar strategy performed with mouseNIH3T3 cells led to the isolation of fisp12,which encodes a secreted 348 residue polypep-tide of 37 kDa.6 The early expression of thesegenes in the cell cycle without the need of priorprotein synthesis categorised them as immedi-ate early genes.

Early studies established that immediateearly genes were expressed transiently and wereeither required for the G0–G1 transition or forG1 progression to S phase. In the first group,c-fos expression is induced between 20 and 90minutes and decreases to undetectable levelsafter two hours, whereas in the second group,c-myc expression is induced between 45minutes and three hours and is detectable up tosix hours after serum stimulation.

Both cyr61 and fisp12 showed the samerapid induction pattern, starting 30 minutesafter serum stimulation and peaking betweenone and two hours after stimulation. In bothcases, short half life (30 minutes for cyr61,10–15 minutes for fisp12) RNA species wereproduced, and sustained expression was ob-served over several hours after the addition ofserum to starved cells. As reported for otherimmediate early genes, the decrease in mRNAvalues requires protein synthesis.

Therefore, it appears that cyr61 and fisp12represent a new class of genes that are induced

Figure 1 The CCN family of genes. The names and chromosomal localisations of thehuman CCN genes are indicated in the left column.

cyr61(1p22.3)

Cystein rich 61 CEF10 (chicken embryo fibroblasts) (Ck)βIG1 (TGFβ induced gene) cyr61 (Mu)

ctgf(6q23.1)

Connective tissuegrowth factor

Ctgf (Ck) ctgf (Sw) ctgf (Xn)bIG2 (TGFβ induced gene) (Mu)Fisp 12 (fibroblast inducible secreted protein) (Mu)

nov(8q24.1)

Nephroblastomaoverexpressed

nov (Ck) novM (Mu) nov (Sw) nov (Xn)

WISP-1(8q24.1–24.3)

Wnt-1 inducedsecreted protein

Elm1 (expressed in low metastatic type 1 cells) (Mu)Wisp 1 (Mu)

WISP-2(20q12–13)


rCop-1 (expression lost after transformation) (Mu)HICP (heparin induced CCN-like protein)Wisp 2 (Mu)

WISP-3(6q23.1)


58 Perbal

www.molpath.com

on Decem


http://mp.bm

j.com/

Mol P


pril 2001. Dow

nloaded from

http://mp.bmj.com/

in the sequential cellular response to mitogenicsignals. The CYR61 and FISP12 proteins arelikely to be required for a relatively long periodduring the progression through the G1 phase.

It is worth pointing out that a second burst ofcyr61 RNA synthesis was seen between six and10 hours after serum stimulation, suggestingthat the expression of this protein might also berequired at a later stage of cell cycle progres-sion (fig 2).

GENES THAT WERE IDENTIFIED ON THE BASIS OF

THEIR INCREASED EXPRESSION IN CELLS TREATED

BY TGFâ1

A considerable number of studies performedwith transforming growth factor â (TGFâ)have established the key role of this signallingmolecule in the regulation of cell growth anddiVerentiation.

For many years, it has been proposed thatthe biological eVects of TGFâ are mediated bya variety of downstream eVectors. To identifygenes in which transcription was aVected byTGFâ1, a cDNA library was prepared fromAKR-2B mouse embryo fibroblasts grown to90% confluency and treated with cyclohex-imide before incubation for six hours in thepresence of 10 ng/ml TGFâ1. DiVerentialscreening of this library with labelled cDNAprobes obtained from untreated and TGFâ1treated cells permitted the cloning of severalgenes whose expression was increased inTGFâ1 treated cells.7 Among them, two struc-turally related genes were designated âIG-M1and âIG-M2. Nucleotide sequencing estab-lished that of âIG-M1 was identical to the pre-viously described cyr61.

Induction of âIG-M1 in growth arrestedcells, as opposed to the induction of CEF10 incells stimulated to divide, provided the firstevidence to suggest that this protein was actingin quite diverse biological processes.7

The sequencing of âIG-M2 revealed that itwas closely related to âIG-M1 in structure.

The âIG-M2 gene was later cloned as ctgf (seebelow).


THEIR ALTERED EXPRESSION IN TUMOURS

Tumour cells are thought to be altered at thelevel of the regulatory signals controllingnormal cell growth. They have provided anunequalled source of biological material inwhich to study the molecular events leading touncontrolled and eventually invasive prolifera-tion.

Isolation of novEarly studies performed in our laboratory hadestablished that the chicken myeloblastosisassociated virus (MAV) induced nephroblas-toma constitutes a unique model of Wilms’stumour, a paediatric tumour aVecting approxi-mately one in 10 000 children.8 Histologically,Wilms’s tumours and the avian nephroblasto-mas are very much alike, with varying amountsof heterotypic diVerentiation occurring, alongwith the development of kidney tumours.9 Forthis reason, Wilms’s tumour is often consideredas a “diVerentiated” tumour.8

MAV belongs to a group of tumorigenic rep-lication competent retroviruses that do notcarry oncogenic sequences of cellular originand usually induce tumours by integrating inthe vicinity of growth regulatory genes in thehost genome.10 Cloning of the proviral integra-tion sites of avian and murine retroviruses ledto the identification of most of the oncogenesand tumour suppressors that turned out to bemajor players in the control of cellular growth.This strategy is particularly useful because itdoes not rely on any particular type of functionfor the target gene.

Because the genetic basis of Wilms’s tumourdevelopment is complex and has not beencompletely elucidated,11 the avian tumoursprovided a unique system in which to identify

Figure 2 Kinetics of CCN gene expression upon stimulation of cell growth.

elm1 nov

2 h 4 h 6 h1 h30 min20 min

fos myc cyr61/ctgf cyr61 cop1

0 24 h8 h

NOV and the CCN family of genes 59

www.molpath.com

on Decem


http://mp.bm

j.com/

Mol P


pril 2001. Dow

nloaded from

http://mp.bmj.com/

the genes underlying the aberrant diVerentia-tion process and the molecular events involvedin the initiation and maintenance of thetumour state. Our analysis of MAV integrationsites enabled us to establish that in one tumourthe MAV proviral genome was integrated intonov,12 a gene sharing structural identity withcyr61, cef10, and ctgf. Based on the observa-tion that the amounts of nov transcripts weresignificantly increased in avian nephroblasto-mas, we assumed that the alteration of nov wasrelated to tumour development. Whether it wasa cause or a consequence still remains to beestablished.

The discovery of nov had three immediateimplications:(1) The structural relation that had been

reported between cyr61 and ctgf wasextended to another gene, thereby leadingto the concept of a new family of genes1 (afamily is usually defined as a group of atleast three subjects sharing common char-acteristics).

(2) nov provided the first example of aberrantexpression of a CCN gene being associatedwith tumour development, therefore indi-cating that the biological properties of thisnew class of growth regulators extendedbeyond the normal response to growthstimulation.

(3) The growth inhibitory eVect of nov onchicken embryonic fibroblasts (CEFs) alsoprovided the first evidence that the CCNfamily of genes contains negative regula-tors and suggested that nov might counter-balance the stimulatory eVects of theimmediate early CCN proteins.

For many years, these last two features didnot receive the attention that they deserved.One possible explanation is that it seemedsomewhat paradoxical that a negative regulatorof growth should be overexpressed in tumourcells prone to active division and proliferation.In fact, the very nature of the nephroblastomaprovides a satisfactory explanation for thisapparent contradiction. As stated above, neph-roblastomas contain an increased number ofdeveloped structures and heterotypic tissues,the diVerentiation of which also requires nega-tive regulatory factors. It is therefore not soheretical that a gene involved in normal diVer-entiation processes (see below) is aberrantlyexpressed in highly diVerentiated tumours.

The inhibition of growth resulting fromretroviral driven expression of nov in secondaryCEFs was not observed upon infection ofBALB/c 3T3 mouse cells with retroviralconstructs expressing the murine orthologue ofnov (P Jolicoeur, 1994, personal communica-tion; EL Suchman et al, 1995, unpublishedresults), therefore suggesting that the growthinhibitory properties of nov might be eVectiveonly in early passage cells, and not in cell lines.Expression of nov was neither detected in qui-escent 3T3 mouse cells, nor in humanfibroblastic cells (S Kyurkchiev and B Perbal,1999, unpublished results).

The location of nov on human chromosome8q24.1, distal to c-myc,13 raised the possibilitythat some pathologies attributed to c-myc

alterations as a consequence of rearrangementsin this chromosome region might also involveaberrant expression of nov. Although oursearch for such cases has not proved successfulas yet, this point certainly deserves more atten-tion and should be the matter of a thoroughanalysis. The recent mapping of WISP-1 to thesame region (see below) reinforces the possi-bility that alterations of CCN genes might playa key role in some pathologies involvingchromosome 8q24.1, previously attributed toalterations of c-myc.

Only recently, the aberrant expression ofCCN genes in tumours and transformed cellshas provided the basis for the identification ofthree new members of the family. Again, it isnot surprising that the various strategies thatwere used led to the isolation of genes involvedat diVerent levels of tumorigenesis and growthcontrol.

Isolation of Elm1In that case, diVerential display methodologywas applied to isolate genes expressed at diVer-ent levels in low (C23) and high (M2)metastatic K-1735 murine metastatic cells.14 Itis important to recall that tumour cell popula-tions are heterogenous with regard to theirability to disseminate and give rise to metas-tases. Low and high metastatic tumour cellsthat are both part of the tumour cell populationare believed to exhibit diVerent biologicalproperties, resulting from the diVerential ex-pression of genes governing fundamental proc-esses such as adhesiveness, spreading, motility,and proliferation.

DiVerential screening of RNA species ex-pressed in the low and high metastatic K-1735variants led to the isolation of Elm1, a geneencoding a 367 amino acid protein, the expres-sion of which was detected in low metastaticbut not high metastatic cells.14 The acquisitionof a metastatic potential is also accompanied byprofound modifications of the post transcrip-tional events that can result in an increasedstabilisation or degradation of RNA tran-scripts. To assess the biological importance ofthe expression of Elm1 being associated with alow metastatic potential, the eVects of ectopicexpression of Elm1 on the tumorigenic poten-tial of high metastatic K-1735 M2 cells werestudied.14 For this purpose, an Elm1 expressionvector was used to transfect and select G418resistant K-1735 M2 derived stable transform-ants expressing diVerent amounts of Elm1RNA, from no expression up to six times theamount of Elm1 expressed in the C23 lowmetastatic variant.

The results obtained indicated that theexpression of Elm1 had pronounced eVects onthe in vivo properties of transfected cells,whereas little eVect was seen on the ex vivogrowth properties of the transfectants, whichonly showed a slight increase of doubling timeand slight decreased saturation density.14

The in vivo eVects were much moreinformative because the expression of Elm1appeared to aVect several parameters.14 First,the absolute number of tumours obtained for agiven period of time was dramatically reduced

60 Perbal

www.molpath.com

on Decem


http://mp.bm

j.com/

Mol P


pril 2001. Dow

nloaded from

http://mp.bmj.com/

by the expression of Elm1. Second, as a resultof the pronounced inhibitory eVect of Elm1 onthe growth properties of transfected cells invivo, the time needed for tumours to reach agiven size was considerably increased. Third,the transfectants that expressed high amountsof Elm1 gave rise to a much smaller number oflung metastases than those expressing low lev-els of Elm1.

It therefore appeared that Elm1 was anotherexample of a CCN protein with a negative typeof activity on cell growth. The induction ofElm1 expression seen three hours after serumstimulation of BALB/c 3T3 cells (fig 2), at amuch later time than cyr61 and ctgf, wouldargue for a role for Elm1 at a relatively latestage of the G1 phase.14

Isolation of rCOP1The identification of genes responsible foroncogenic transformation of normal cells hasbeen the subject of many studies over the past20 years. Despite the considerable amount ofinformation that has emerged from these stud-ies, most of the molecular events that governprogression from the normal to immortalisedand transformed stages remain to be uncov-ered. Thus, the functional bases for thecooperation between the p53 tumour suppres-sor and activated H-ras oncogene in inducingoncogenic transformation are still obscure.

The diVerential display strategy that wasused to screen mRNAs expressed in normal ratembryo fibroblasts and their counterpartstransformed by p53 mutants in cooperationwith activated H-ras has allowed the identifica-tion of rCOP1, a new gene of the CCN family,encoding a 250 amino acid protein, the expres-sion of which was downregulated in trans-formed cells.15

The expression of rCOP1 was detected inmouse BALB/c 3T3 cells but was repressed inseveral transformed derivatives, and retroviraldriven ectopic expression of rCOP1 estab-lished that rCOP1 expression had a strongnegative eVect on the growth of rat trans-formed cells, leading to a cell population show-ing enrichment in sub-G1 DNA content. Fur-thermore, rCOP1 overexpression in stablytransfected rat cells transformed by p53 andH-ras significantly reduced their tumorigenic-ity when tested by subcutaneous injection inathymic mice.

These results15 identified rCOP1 as the thirdmember of the CCN family to exhibit proper-ties of a negative regulator of growth.

Interestingly, rCOP1 expression was de-tected only after repeated passages of rat andmouse embryo fibroblasts and not in primaryfibroblasts.15 In addition, adult rat tissues didnot appear to contain any rCOP1 RNAspecies, indicating that the expression ofrCOP1 was not induced as the result of growtharrest in non-proliferating, fully diVerentiatedtissues.

This is another example of the expression ofsome CCN genes being susceptible to eventsleading to the “establishment” of cell lines (seeabove for nov) and stresses the fact that careshould be taken when comparing patterns of

CCN gene expression obtained in embryonicfibroblasts and established cell lines.

Isolation of WISP genesThe suppressive subtractive hybridisation(SSH) strategy was used to identify targets ofthe Wnt-1 protein by screening RNAs diVeren-tially expressed in C57MG mouse mammaryepithelial cells and their transformed counter-part obtained after infection with a wnt-1expressing retrovirus.16 The WNT-1 proteinwas originally isolated as an insertion site formouse mammary tumour virus in murinemammary tumour virus (MMTV) inducedmammary adenocarcinomas. It is a member ofan expanding family of signalling proteinsinvolved in the control of normal cell prolifera-tion and in tumorigenesis.

Two genes (WISP1 and WISP2),16 theexpression of which was upregulated in wnt-1transformed cells, were shown to be identical tothe previously described Elm1 and rCOP1genes, respectively. Their expression was notincreased by overexpression of wnt-4, whichdoes not induce morphological transformationof the C57MG cells, therefore suggesting thatWISP expression might play a role in the trans-formation process. The WISP-1 gene has beenmapped to chromosome 8q24.1–8q24.3 in thevicinity of nov, whereas the WISP-2 gene hasbeen mapped to chromosome 20q12–20q13.1,a region frequently amplified in node negativebreast cancers with poor prognosis and inmany colon cancers.

The expression patterns and biologicalproperties of WISP1 and WISP2 have bothreinforced the idea that the biological proper-ties of CCN proteins might be dependent uponthe cellular context and have pointed out thecomplexity of the regulatory circuitry in whichthese proteins function.

For example, the upregulation of rCOP1/WISP-2 in wnt-1 transformed C57MG epithe-lial cells16 was in apparent contradiction to thedownregulation of this gene upon transforma-tion of BALB/c 3T3 fibroblastic cells.15 Thissuggests that the CCN proteins may haveopposing functions in diVerent cell types. Fur-thermore, in primary human colon adenocarci-nomas, the expression of WISP-2 was signifi-cantly decreased, and it was not detected in theepithelial tumour cells of mammary carcinomaobtained from wnt-1 transgenic mice.16 In theWISP positive stroma, spindle shaped cellsadjacent to capillary vessels stained positive forWISP-2 expression in these tumours.

Similarly, the expression of Elm1/WISP-1,which had been reported to suppress the meta-static potential of murine melanoma cells,14

was significantly increased in most humancolon carcinomas.16 Increased expression ofWISP-1 was seen in the stroma of mousemammary carcinomas produced in wnt-1transgenic mice (which stained strongly posi-tive for WISP-1 expression), whereas tumourepithelial cells expressed low amounts ofWISP-1.16

In apparent contradiction to the tumoursuppressor type of activity assigned to Elm1,14

overexpression of WISP1 induced accelerated


www.molpath.com

on Decem


http://mp.bm

j.com/

Mol P


pril 2001. Dow

nloaded from

http://mp.bmj.com/

growth and morphological transformation ofNRK-49F fibroblastic cells.16 Furthermore,expression of WISP-1 was found to increasethe tumorigenicity of two NRK-49F cell linesestablished after infection by retrovirus con-structs expressing WISP-1. Because these cellswould not grow without anchorage, the overex-pression of WISP-1 was not suYcient to allevi-ate the need for proper interactions with theextracellular matrix and attachment.16


THEIR DIFFERENTIAL EXPRESSION IN RSV

TRANSFORMED FIBROBLASTS

Transformation of CEFs by pp60v-src of Roussarcoma virus (RSV) is known to alter, eitherpositively or negatively, the transcription of avariety of regulatory and signalling proteins.Because the expression of pp60v-src had beenreported to block the diVerentiation of myo-genic cells, it was proposed that genes whoseexpression is decreased by pp60v-src transforma-tion might act to maintain or induce the diVer-entiated state of target cells.

The strategy used to characterise genesacting in the transduction pathways activatedfor transformation was usually based on theuse of CEFs transformed by a mutant ofRSV (tsNY72) expressing a thermosensitivepp60v-src. Upon shifting to the permissivetemperature (35°C), the constitutive tyrosinekinase activity of pp60v-src induces the cascadeof molecular events that lead to oncogenictransformation.

In the first set of experiments, a cDNAlibrary was prepared from tsNY72 infectedcells maintained at the permissive temperaturefor four to six hours in the presence of 75 µMcycloheximide, and screened with a subtractedprobe made of labelled cDNA from tsNY72infected cells grown at 35°C hybridised againstmRNA from unstimulated cells. Clones ob-tained with this strategy corresponded to stabi-lised mRNA whose expression had beeninduced by pp60v-src. Twelve diVerent immedi-ate early genes were induced, includingCEF10, the chicken equivalent of cyr61.17

Upon serum stimulation of starved CEFs,the pattern of CEF10 expression was similar tothat of cyr61 in murine cells (fig 2). However,no late burst of induction was reported forCEF10. Whether this discrepancy results fromvariations in the protocols and origin of cellsused or represents important diVerences in thebiological properties of murine and chickengenes is not yet clear.

The pattern of CEF10 expression in pp60v-src

induced cells is interesting. In pp60v-src inducedcells, CEF10 expression was found to be mul-tiphasic, with two peaks at one and eight hourspost shift, a decrease at 12 hours post shift, anda third peak at 24 hours.17 Under similarconditions, most immediate early genes werepersistently induced for 24 hours. When amitogenic non-transforming myristoylation de-fective virus NY315 was used, CEF10 wasinduced in a linear mode for a period of 24hours.17 These observations indicated thatCEF10 production was increased as a result ofthe pp60v-src growth stimulatory eVect. The

cyclic expression of CEF10, which was notobserved with NY315, might be related to thetransformation process induced by v-src.

In another set of experiments, the diVerentialdisplay strategy was applied to cDNAs pre-pared from tsNY72 RSV infected CEFs grownat the non-permissive temperature and eithershifted to the permissive temperature or main-tained at the non-permissive temperature.Screening of the genes diVerentially expressedin these cells revealed that the expression of thenov gene was downregulated upon the stimula-tion of CEF growth induced by pp60v-src.18

Nuclear run on assays performed on serumstarved uninfected CEFs and wild-typepp60v-src transformed CEFs established that thediVerential rate of nov transcription in isolatednuclei did not account for the diVerence in thesteady state of nov transcripts measured in qui-escent and transformed CEFs. Therefore, theincrease levels detected in uninfected cellsmight result from an accumulation of the novtranscripts as a result of the long half life (eighthours) of the nov transcripts in quiescent cells.

The use of a mutant strain expressing a ther-mosensitive pp60v-src revealed that upon shiftingto the non-permissive temperature, there was adecline in the numbers of nov transcriptsdetected in transformed cells between four andeight hours, with a dramatic reduction after 12hours, and an absence of transcripts after 24hours.18 A similar decline in nov expression wasseen when CEFs were stimulated to proliferateeither after transformation by various onco-genes or after infection with the non-transforming myristoylation defective NY315mutant.18 Therefore, the downregulation ofnov expression induced by the stimulation ofcell growth and the high concentrations ofNOV that were detected in quiescent cells werein striking contrast to the situation observed forother CCN genes, and established that nov wasnot an immediate early gene (fig 2). Therequirement for protein and RNA synthesis inthe decrease of nov transcripts upon stimula-tion of cellular growth indicated that thereduced expression of nov resulted either fromthe synthesis of a negative regulator of novtranscription or from an increased turnover ofnov transcripts.18

Our observation that the expression of novwas modulated during the progression of thecell cycle and reached a peak at the G2 phase(S Middendorp and B Perbal, 1995, unpub-lished results) suggests that the increasedsynthesis and accumulation of stable nov tran-scripts at that late stage might be required forproper diVerentiation.

GENES THAT WERE ISOLATED ON THE BASIS OF AN

IMMUNOLOGICAL CROSSREACTIVITY OR A

STRUCTURAL IDENTITY WITH CCN GENES

Polyclonal antibodies raised against plateletderived growth factor (PDGF) were used tocharacterise and isolate the protein(s) responsi-ble for the PDGF related mioattractant activitythat was detected in conditioned medium fromhuman umbilical vein endothelial cells (HU-VEC).19 This approach has led to the identifi-cation of an anti-PDGF crossreacting protein

62 Perbal

www.molpath.com

on Decem


http://mp.bm

j.com/

Mol P


pril 2001. Dow

nloaded from

http://mp.bmj.com/

with a reported apparent molecular mass of38 kDa. This protein, which was shown to pos-sess the major chemotactic and mitogenicactivity previously attributed to PDGF, wasdesignated CTGF (connective tissue growthfactor).19 Cloning of CTGF revealed that it wasthe human orthologue of fisp12 and âIG-M2(see above). As expected from its structuralidentity with fisp12, the ctgf gene was shown tobe rapidly induced by serum in human skinfibroblasts (fig 2), independently of proteinsynthesis, therefore confirming that it wasindeed an immediate early gene.20

In agreement with previous studies that ledto the discovery of âIG-M2, the ctgf gene,which is expressed in unstimulated fibro-blasts,21 22 was reported to be induced by TGFâin the absence of protein synthesis.7 20 23 Fromthese observations, a potential role for CTGFin fibrotic diseases has progressively emerged,and it is now well established that ctgf mRNAand protein are overexpressed in fibroticlesions of several major organs in humans,including liver, kidney, gingiva, cardiovascularsystem, pancreas, bowel, and eye (reviewed byBrigstock3 and Essam and colleagues4). How-ever, the relation between TGFâ and ctgf is notobligatory because several studies have shownthat the activation of ctgf expression canproceed independently of TGFâ (reviewed byEssam and colleagues4).

Because of its identification as a downstreamtarget of TGFâ, ctgf has been studied exten-sively over the past few years and our growingknowledge indicates that, as reported for otherCCN genes, its biological properties and regu-lation of expression are dependent upon thecellular context and the nature of the cells inwhich it is produced and/or acting. Initiallyreported to induce cell proliferation, survival,and extracellular matrix production, CTGFhas also been shown to be apoptotic in humanaortic smooth muscle cells and human breastcancer MCF-7 cells.24–26

Another experimental approach, based onthe screening of expressed sequence tag (EST)databases, with the WISP-1 protein sequence,16

allowed the identification of WISP-3, a geneencoding a 354 amino acid new member of theCCN family with slightly diVerent structuralfeatures to the other members: it is the onlymember thus far that does not show thecharacteristic conservation of cysteines.

Until recently, very little information wasavailable about the biological properties ofWISP-3. The WISP-3 gene has been reportedto be significantly overexpressed in humancolon carcinoma,16 and putative loss of func-tion mutations in the WISP-3 gene have beenassociated with progressive pseudorheumatoiddysplasia in humans, an autosomal recessiveform of spondyloepiphyseal dysplasia.27

On the one hand, the diVerent systems inwhich the CCN genes have been identifiedmade it obvious that their expression ought tobe involved in a wide variety of biological proc-esses in normal and pathological conditions.However, on the other hand, the strikingconservation of their modular organisationwith 38 cysteins, the position and number of

which have been conserved throughout thefamily members and during evolution, sug-gested that they might share some kind offunctionality.

Until now, most of the studies performedwith the various CCN genes and proteins havebeen descriptive and have been performed ondiVerent biological systems. Unfortunately,only in a very few instances, were studiesperformed in parallel to compare the expres-sion or the activity of diVerent members of thefamily. This type of approach might turn out tobe very fruitful in the case of CCN proteins forwhich no biological assay is available as yet.Obtaining this type of information throughcollaborative eVorts would undoubtedly lead toan integrated view of CCN gene expressionand functions at a very reduced cost.

The use of diVerent systems has increasedthe number of results that appear to be“conflicting” if one does not take into accountthe variety of physiological conditions consid-ered. There is no doubt that if the CCNproteins play, as one can guess, key roles in somany diVerent fundamental regulatory proc-esses, the major issue is understanding thestructural basis for their biological functions.This goal will be reached by the combination ofdiVerent techniques, including purification ofbiologically active CCN proteins, identificationof their targets and cofactors, directed muta-genesis of crucial amino acids, the constructionof chimaeras, and the manipulation of genes inthe context of living animals.

Before this goal can be achieved, there areissues that remain unresolved and need to beconsidered for each member in the context ofthe CCN family as a whole.

Based on the results that we have obtainedwith nov I will consider below: (1) the relationbetween the structure and subcellular localisa-tion of CCN proteins, (2) the mechanismsgoverning the temporal and spatial regulationof CCN gene expression, (3) the implication ofthe CCN proteins in pathological conditions,and (4) the identification of proteins andligands that interact with CCN proteins.

Structure and subcellular localisation ofCCN proteinsFigure 3 shows the modular structure of theCCN proteins. Although they have a very con-served multimodular organisation, with fourmodules sharing identity with insulin-likegrowth factor binding proteins (IGFBPs), VonWillebrand factor, thrombospondin, and acystein knot containing family of growth regu-lators, the CCN proteins have distinctivebiological properties and are diVerentiallyregulated.

Because of their partial identity withIGFBPs and the low aYnity binding of IGF-Iand IGF-II to CTGF,28 a few groups in the IGFfield have attempted to rename the CCN pro-teins using an IGFBP related system. Althoughit is quite possible that both IGFBPs and CCNproteins participate in common regulatorypathways, there is as yet no functional evidenceto support such a renaming.29 From discus-sions at the first international workshop on the


www.molpath.com

on Decem


http://mp.bm

j.com/

Mol P


pril 2001. Dow

nloaded from

http://mp.bmj.com/

CCN family of genes, a unifying nomenclatureis now proposed for the CCN genes andproteins (table 1). It is also suggested that thenomenclature committee of the CCN society([email protected]) should be con-tacted for the attribution of names for newCCN members.

The multimodular organisation of the CCNproteins raises interesting questions as to theparticipation of each module in the function ofthe full length protein. Either the activities ofeach module add up or they confer on thewhole protein specific functions that mightsubstitute or add to the function of theindividual modules. Such a modular structurealso suggests that the diVerent modules corres-pond to functional domains that can interacteither sequentially or simultaneously withother partners, and that the final biologicalproperties of the CCN proteins might bedependent upon diVerent combinatorial ef-fects.

Studies performed with NOV, CTGF, andWISP-2 have provided valuable informationabout the structural basis for some biologicalproperties of the CCN proteins.

The nov gene was discovered as an integra-tion site for the MAV retrovirus in avian neph-roblastomas (see above).12 This gene wasdesignated nov because it was highly expressedin all nephroblastomas compared with adultkidney, where its expression was low but stilldetectable.9 In one sample (tumour 725), theintegration of MAV into the second intron ofnov resulted in the U5 long terminal repeat(LTR) driven expression of a chimaeric novmRNA containing 91 3'-proximal viralnucleotides and 2.0 kb of cellular sequences.As a consequence, the first putative initiationcodon corresponded to an internal methioninecodon in the full length NOV protein, and theresulting chimaeric protein, composed of 253amino acids, was expected not to be secretedbecause it was deprived of the signal peptide

identified at the N-terminus of NOV and allother CCN proteins. Although increased ex-pression of nov was detected in all nephroblas-tomas tested (including those induced by theMAV2-O strain, which also induces osteopet-rosis), none of them was found to contain aphysically altered nov gene (CL Li et al, 2000,unpublished results).12

The construction of retroviral competentovian recombinants, expressing either the fulllength nov RNA or the truncated version ofnov expressed in tumour 725, allowed us toestablish that the ectopic overexpression of thefull length nov in actively growing CEFs(which did not express endogenous nov) inhib-ited their growth.12 Recombinant viruses repli-cated in these cells (as shown by RNA expres-sion) and gave rise to viral stocks; the growthinhibitory eVect of nov did not induce celldeath but rather blocked the cells in the cellcycle, probably at the G2–M transition. On thecontrary, the expression of the truncated novRNA species did not interfere with the growthof the CEFs and induced their morphologicaltransformation.12 From these results it wasconcluded that nov was a protooncogene withantiproliferative activity. Therefore, it could beconcluded that the development of nephroblas-tomas in chickens was associated with anincreased expression of either a full length or atruncated NOV protein.

Because the fates of truncated and full lengthNOV proteins are expected to be quitediVerent (fig 4), these results suggested that thebiological properties of the NOV proteinsmight depend upon their structure and theirsubcellular localisation.

The immunofluorescence detection of aNOV related protein in the nucleus of severaldiVerent cell lines and frozen tissue prepara-tions raised the intriguing possibility that itmight be involved more directly in theregulation of gene expression. Because theamount of nuclear NOV protein was found tovary greatly (from zero to high concentrations)among various cell lines (B Perbal, 1997,unpublished observations and PN Schofield,1998, personal communication), the nuclearlocalisation of NOV might be related to theirproliferation and cell cycle status.

Several lines of evidence confirmed theexistence of a nuclear NOV related protein and

Figure 3 Multimodular structure of the CCN proteins. CT, cystein knot containing familyof growth regulators-like domain; IGFBP, insulin-like growth factor binding protein-likedomain; TSP1, thrombospondin-like domain; and VWC, Von Willebrand factor-likedomain.

CTGF 100/100

%ID/HO IGFBP VWC TSP1 CT

CYR61 44/60

NOV 48/64

ELM1/WISP1 41/57

COPI/WISP2 32/41

WISP3 36/52

Variable

Variable

Variable

Variable

Variable

Variable

Table 1 Proposed nomenclature for CCN proteins

CCN nomenclature Current names Alternative designations

CCN 1 CYR61/CCE10 IGFBP-rP 4CCN 2 CTGF/FISP12 IGFBP-rP 2, Hcs 24CCN 3 NOV IGFBP-rP 3CCN 4 WISP-1/ELM 1CCN 5 WISP-2/rCOP-1CCN 6 WISP-3

CCN names were given according to their chronological identification. It is suggested that a lowercase letter should be used to indicate species when needed (for example, hCCN1 for humancyr61).

Figure 4 Fate of the full length and truncated proteins.Rous sarcoma virus (RSV) driven expression of thetruncated protein in chicken embryonic fibroblasts (CEFs)leads to the cytoplasmic accumulation (and furtherprocessing?) of the truncated NOV protein, whereas thepresence of the signal peptide in full length NOV governs itssecretion.

RSV Full length nov

RSV 5' truncated nov

64 Perbal

www.molpath.com

on Decem


http://mp.bm

j.com/

Mol P


pril 2001. Dow

nloaded from

http://mp.bmj.com/

suggested that it was indeed involved in theregulation of transcription.

AN AMINO TRUNCATED NOV RELATED PROTEIN IS

LOCALISED IN THE NUCLEUS OF HeLa AND 143

CELLS

A 31/32 kDa doublet of NOV protein wasidentified by western blotting in the nuclearfraction of human epitheloid carcinoma HeLacells, and human osteosarcoma 143 cells.30

Because the antibodies used in this study weredirected against the C-terminus of NOV, the31/32 kDa NOV isoform was thought to be anamino truncated form of NOV.

The nuclear detection of NOV was reminis-cent of a situation reported previously for sev-eral other secretory proteins. In the absence ofa conventional nuclear localisation signal(NLS) in NOV, its nuclearisation might begoverned by the KKGKKCLRTKKS motif,which is similar to the basic motif involved inthe nuclear localisation of fibroblast growthfactor 3 (FGF3).

Although the biological importance of NOVnuclear targeting remains unclear, our expecta-tion is that the 31/32 kDa NOV isoform mightmodulate gene expression during normaldevelopment. The biochemical identificationof this nuclear NOV related protein is underway.

THE NUCLEAR 31/32 kDa NOV RELATED PROTEIN

COLOCALISES WITH THE TRANSCRIPTIONAL

MACHINERY IN THE 143 CELLS

As a first step in establishing the localisation ofthe nov protein in the nucleus of these cells, wetook advantage of the fact that herpes simplexvirus 1 (HSV-1) encoded infected cell proteins4 and 8 (ICP4 and ICP8) can be used asphysical markers for the transcription and rep-lication machineries, respectively. The ICP4protein is a transcriptional transactivator local-ised in the nucleus of HSV infected cells, whereit is responsible for the regulation of HSV geneexpression, whereas ICP8 is a single strandedDNA binding protein that is found in thenuclear compartment, where viral DNA syn-thesis takes place.31

To determine whether NOV protein couldbe detected in the nucleus of actively growing

cells, three series of experiments were per-formed. In the first, uninfected fixed 143 cellswere exposed to the anti-nov K19 rabbitantibody and then to biotinylated goat antirab-bit immunoglobulin, followed by avidin boundto Texas red. In parallel experiments, cells weretreated with the anti-nov K19 antibody andincubated directly with goat antirabbit Texasred conjugated antibodies. Nuclei from unin-fected 143 (fig 5A) cells showed a strong posi-tive response to the K19 anti-nov antibody. Inuninfected actively growing cells, the nuclearNOV protein appeared to be localised in thenucleoplasm and in nucleoli. Positive stainingwas also seen at the periphery of the nucleus(probably in the Golgi) and in the cytoplasm ofthese cells.

In the second series of experiments, cellswere infected with HSV-1 before incubationwith rabbit anti-nov (K19), mouse anti-ICP4,and mouse anti-ICP8 antibodies. It has beenreported previously that early after HSV infec-tion, ICP4 is diVusely distributed throughoutthe infected cell nucleus and that, after theonset of viral DNA replication, staining forICP4 was associated with discrete densebodies. In contrast, ICP8 labelling was fairlyconstant throughout infection, giving rise to agranular pattern that filled the infected nu-cleus.

Infected 143 cells were first incubatedsimultaneously with either both the rabbitanti-K19 and the mouse anti-ICP4 antibodiesand then with goat antimouse fluoresceinisothiocyanate (FITC) conjugated antibodies(reacting with anti-ICP4 and anti-ICP8) orwith biotinylated goat antirabbit immuno-globulin before incubation in the presence ofTexas red conjugated avidin (allowing indirectvisualisation of the K19 antibody).

As shown in fig 5B, ICP4 (green fluores-cence) and NOV (red fluorescence) overlappedin the nucleus of infected 143 cells, whereasICP8 (green fluorescence) and NOV (red fluo-rescence) did not colocalise (fig 5C). Asexpected, the nucleoli did not contain dualstaining. These experiments suggested that themajor fraction of the nuclear NOV isoform waslocalised in the compartment where transcrip-tion occurs and was not associated with the

Figure 5 Confocal immunocytolocalisation of NOV in the nucleus of 143 osteosarcoma cells. (A) staining with anti-nov antibodies, (B) double stainingwith anti-nov and anti-HSV-ICP4 (herpes simplex virus infected cell protein 4) antibodies, (C) double staining with anti-nov and HSV-ICP8 antibodies.The yellow colour indicates colocalisation of ICP4 and NOV.


www.molpath.com

on Decem


http://mp.bm

j.com/

Mol P


pril 2001. Dow

nloaded from

http://mp.bmj.com/

replication machinery (B Perbal, 1997, unpub-lished data).

THE NOV PROTEIN INTERACTS WITH THE RPB7

SUBUNIT OF RNA POLYMERASE IN A YEAST TWO

HYBRID SYSTEM

A yeast two hybrid screening system was usedto isolate cDNAs encoding proteins able tointeract with the human NOV protein.30 ADNA fragment encoding NOV codons 29 to357 was amplified by the polymerase chainreaction (PCR) and inserted into pGBT9(Clontech, Palo Alto, California, USA) inframe with the DNA binding domain of Gal4.The resulting plasmid (pNOV) was used as“bait” to screen three cDNA libraries (anEpstein-Barr virus transformed human pe-ripheral blood lymphocyte library (gift fromAviron Inc.), a HeLa cell library, and a normalhuman brain library from Clontech, allprovided by Dr B Roizman, Kovler Labora-tory, University of Chicago, USA), which werefused to the Gal4 transcriptional activationdomain in pACT.

Establishing the 3'-proximal sequence of theinserts at the 5' boundary of the Gal4 DNAdomain fusion (B Perbal, 1997, unpublishedresults; CP Li et al, 2000, unpublishedresults)30 allowed us to identify several inde-pendent clones with sequences that matchcompletely with the published sequence of therpb7 subunit of human RNA polymerase II.Because it has been established that the rpb7subunit alone does not give rise to falsepositives under the conditions used (MVerner, 1998, personal communication), ourresults strongly suggested that the nuclearamino truncated NOV related protein isinvolved in the regulation of transcriptionthrough its interaction with the RNA polymer-ase holoenzyme.

A NOV RELATED PROTEIN CAN BE DETECTED AT

THE NUCLEAR ENVELOPE

We have shown that the NCI-H295R cellsthat have been established from an invasiveadrenocortical carcinoma secrete increasedconcentrations of NOV protein (S Kyurkchievet al, Proceedings of the first internationalworkshop on the CCN family of genes, 17–19October 2000, Saint-Malo, France). When theimmunogold procedure was used to performan immunocytochemical localisation of theNOV protein in these cells, we found that thecytoplasm, the plasma membrane, and thenuclear envelope stained positive. Unexpect-edly, the distribution of the positive grains inthe cytoplasm of the NCI-H295R cells and theabsence of any recognisable secretory granuleor secretory vesicle suggested that shuttlingof the NOV protein in these cells didnot follow the conventional pathway (GThomopoulos et al, 2001, in press). The modeof secretion of the protein is currently underinvestigation.

The detection of an immunoreactive NOVrelated protein at the nuclear pores of NCI-H295R cells was another piece of evidencesuggesting that NOV might be targeted to thecell nucleus.

A NOV RELATED PROTEIN BINDS THE PROMOTER

OF HUMAN PLASMINOGEN ACTIVATOR INHIBITOR

TYPE 2

In recent studies performed to identify the fac-tors responsible for the transcriptional regula-tion of the human plasminogen activatorinhibitor type 2 (PAI-2) gene, competitionelectrophoretic mobility shift assays identifiedtwo cDNAs from a HeLa cell library encodingpartial peptides that specifically bound to thetranscriptional regulatory motif (TRM), previ-ously found to be essential for constitutivePAI-2 transcription. Sequencing of the cDNAsestablished that one of these peptides wasidentical to the fifth domain of NOV,32

therefore reinforcing the possibility that a NOVrelated protein might indeed be involved intranscriptional regulation.

Because there is no evidence as yet for alter-native splicing of the nov RNA, the detection ofa NOV related protein in the nucleus and in thecytoplasm of the cells raises fundamental ques-tions as to the biochemical constitution of theseproteins and their association. To investigatethese questions, we have now constructeddiVerent types of vectors expressing tagged fulllength and truncated NOV proteins. Studiesaimed at characterising processes governingthe turnover and subcellular localisation of thediVerent NOV related proteins and their func-tional relation have been undertaken.

In the absence of compelling evidence for theexpression of new members of the CCN familyexhibiting a nuclear localisation, it is temptingto relate these various observations to thedetection of an amino truncated NOV isoformin the conditioned medium of nov expressingcells.

The analysis of conditioned medium fromSF9 insect cells infected with a nov expressingrecombinant baculovirus identified a 25 kDaNOV protein as an amino truncated isoform.N-terminal sequencing performed on the puri-fied secreted protein revealed that the trun-cated protein was made of the two C-proximaldomains of NOV, with an N-terminal sequence(AYRPE), suggesting that it was the result of aproteolytic cleavage of the full length secretedNOV protein.33 The existence of a truncatedform of CTGF with a similar N-terminalsequence34 suggests that specific proteolysis ofthese CCN proteins might be a key element inthe regulation of their biological activity.

A truncated NOV protein with an apparentmolecular mass of 30–32 kDa can be detectedin the conditioned medium of naturally ex-pressing cells such as NCI-H295R (GThomopoulos et al, 2001, in press); transfectedMadin Darby canine kidney (MDCK) cells35;transfected human glioblastomas (B Perbal,2000, unpublished results); in five day oldhuman myoblasts, but not in myotubes (SKyurkchiev and B Perbal, 1999, unpublishedresults); and in diVerent human biological flu-ids, such as urine, amniotic fluid,36 andcerebrospinal fluid.36 Our working hypothesis isbased on the assumption that the eliminationof the N-proximal modules confer on the trun-cated NOV protein growth stimulatory proper-ties, as opposed to the inhibitory properties of

66 Perbal

www.molpath.com

on Decem


http://mp.bm

j.com/

Mol P


pril 2001. Dow

nloaded from

http://mp.bmj.com/

the full length NOV protein (see above). Theoncogenic properties of the truncated NOVprotein expressed in avian nephroblastoma (seeabove), and the growth inhibitory properties ofrCOP1,15 might therefore result from thedisruption of the regulatory balance existing inthe full length protein, between a negativeN-proximal domain and a positive C-terminaldomain. In the case of an amino truncatedNOV protein, constitutive positive propertieswould result from the lack of negative regula-tion conferred by the first module, whereas inthe case of rCOP1, the lack of positive eVectswould result from the lost of the C-terminalmodule.

Indirect evidence suggesting that theC-terminus of NOV is not freely accessible inthe native protein came from studies aimed atsetting up an enzyme linked immunosorbentassay (ELISA) to measure the concentrationsof NOV in biological fluids (S Kyurkchiev et al,Proceedings of the first international workshopon the CCN family of genes, 17–19 October2000, Saint-Malo, France). By performingELISA with phosphate buVered saline (PBS)or Triton treated NCI-H295R cells we estab-lished that the NOV protein was present at thecell membrane and that a considerable amountof NOV could be found in the cytoplasm ofthese cells. This conclusion was confirmed byindirect immunofluorescence assays, whichindicated that the NOV protein was widely dis-tributed in the cytoplasm of the NCI-H295Rcells, sometimes associated with perinucleargranules, suggesting labelling of the Golgi.Unexpectedly, the NOV protein was notdetected when absorption ELISA was per-formed with conditioned medium from NCI-H295R cells in which the presence of secretedNOV had been checked for by westernblotting. This suggests that in the native NOV

protein, the C-terminus of NOV was notaccessible to the antibodies, either because ofconformational constraints or because it wascombined with other partners.

The possibility that physical interactionsbetween the C-proximal and N-proximaldomains of NOV can regulate the activity ofthe native protein is under current investiga-tion.

Thus, the balanced production of truncatedand full length NOV isoforms might be tightlyregulated and be dependent upon physiologicalconditions. This aspect might be particularlyimportant in the case of cancer cells, whichoverproduce many diVerent kinds of proteases.The disregulation of proteolysis would increasethe relative ratio of truncated over full lengthisoforms, therefore leading to abnormal signal-ling. However, the situation is probably not thatsimple. In the course of the studies that we haveinitiated to identify the molecular eventsleading to the secretion of NOV isoforms, wehave established that the concentration oftruncated NOV protein in the conditionedmedium of eukaryotic cells is not relateddirectly to the amount of the released fulllength 48 kDa protein (V Martinez and B Per-bal, 2000, unpublished results), therefore sug-gesting that the production of the amino trun-cated NOV isoform is not the sole result ofNOV post translational processing.

The possibility that the detection of anuclear NOV protein did not result from theintracellular addressing of NOV, but wasrelated to the secretion of a full length NOV,was suggested by experiments in which inhibi-tion of the secretion of the full length NOV didnot increase the concentration of nuclear NOVin the treated cells (B Perbal, 1997, unpub-lished observations). It is tempting to speculatethat the truncated NOV protein is targeted tothe nucleus after post translational proteolysisof the secreted NOV protein, either at the sur-face of the cell or in the matrix by one of themechanisms depicted in fig 6.

The isolation of several biologically activetruncated forms of CTGF in biological flu-ids3 34 is another example that highlights thepotential importance of post translationalmodifications of CCN proteins, and stressesthe need for tight control to ensure balancedproduction of the various isoforms if process-ing and biological activity are related. In thatcase, the levels of low mass stable CTGFisoforms were proposed to be qualitatively andquantitatively dependent upon several factors,such as tissue of origin, species, or stage of thecell cycle.3 Again, the amino truncated 10 kDaCTGF, which contains only the C-terminalmodule of CCN proteins, exhibited mitogenicproperties, confirming that the N-proximalmodules are not required for the positive regu-lation of growth.3 However, the complexity ofthe structure–activity relation is exemplified bythe fact that the 10 kDa CTGF molecule ismitogenic for BALB/c 3T3 cells, vascularsmooth muscle cells, and endometrial stromalcells but not endothelial cells.3

The study of the mechanisms governing theproduction of the truncated CCN proteins

Figure 6 Schematic hypothetical pathways leading to the potential internalisation of theamino truncated NOV protein. In all cases, the secretion of a full length NOV protein isthought to be required for further processing. The full length NOV protein might be secretedand processed at the membrane site, the C-proximal moiety of the protein being internaliseddirectly or indirectly through specific interactions with either a receptor or a transporter.Alternatively, the NOV protein could be released into the medium or the extracellularmatrix where it would be submitted to proteolysis. The amino part of the protein might itselfplay an important role in signalling, either directly or indirectly. C-ter, amino truncatedprotein; N-ter, N-proximal moiety of the protein; R, receptor; T, transporter.

N-ter

N-ter

N-ter

C-ter

N-ter

C-ter

N-ter

C-ter

C-ter C-ter

R

T

T

C-ter

C-ter

C-ter

Nucleus

Cytoplasm

C-ter


www.molpath.com

on Decem


http://mp.bm

j.com/

Mol P


pril 2001. Dow

nloaded from

http://mp.bmj.com/

should result in a better insight into thebiochemical properties of the four constitutivemodules and their functional interactions.

Regulation of CCN RNA and proteinexpressionAnother fundamental question regarding thebiological properties of the CCN proteins con-cerns the way their expression is regulated bothin space and time. Because a large quantity ofinformation concerning the patterns of cyr61and ctgf expression has been published inrecent reviews,2–4 I will focus on resultsobtained with nov and will attempt to highlightthe specificities of each system.

PROMOTER SEQUENCES AND GENE EXPRESSION

As stated above, the expression of the cyr61and ctgf genes is induced early upon stimula-tion of fibroblast proliferation, whereas Elm1and rCOP-1 exhibit delayed kinetics of induc-tion and the expression of nov is downregu-lated under similar conditions (fig 2). This dif-ferential expression pattern and the variety ofeVectors involved in the activation or repres-sion of CCN genes have suggested that distinctregulatory elements govern their expression.Indeed, the comparison of ctgf, cyr61, and novpromoter sequences did not show any obviousconservation.3 37–39

The induction of cyr61 by serum and PDGFin fibroblasts was reported to be dependentupon a serum responsive element (SRE)-likemotif at position −1912/−1933 in the 2 kb5'-proximal promoter sequences, which aresuYcient to confer serum inducibility.37 Theexpression of ctgf was also found to bestimulated by PDGF and epidermal growthfactor (EGF) in human fibroblasts.23 Similarly,both cyr61 and ctgf were reported to beinduced by fibroblast growth factor (FGF).5 23

The expression of cyr61 is also induced byvitamin D3 in human fetal osteoblasts,40 byoestrogen and tamoxifen in uterine cells ofovariectomised rats,41 and hippocampal im-mortalised cells undergoing neuronal diVeren-tiation.42

The rapid induction of cyr61 (peak after 20minutes, decreased at 60 minutes, and absentat two hours) by 100 ng/ml 12-O-tetradecanoylphorbol-13-acetate (TPA)5 wasin pronounced contrast to the expression ofCEF10 (the avian homologue of cyr61), whichwas reported to be repressed to < 20% of thebasal value at 60 minutes, and graduallyincreased to about 10 times the basal value at24 hours.17 Based on the time course for thetranslocation–activation–deactivation of pro-tein kinase C (PKC), the expression pattern ofCEF10 in the presence of 100 ng/ml TPA sug-gested that the expression of CEF10 isrepressed at one hour, when PKC is transientlyactivated, and derepressed when PKC isdegraded.17 Similarly, stimulation of serumstarved cells by 50 ng/ml TPA for six hoursresulted in the complete downregulation of novin CEFs.18 The downregulation of nov expres-sion induced by serum and by TPA could beabolished by treatment with the H7 inhibitor ofPKC, suggesting that PKC might play a role in

the decrease in the half life of nov RNAtranscripts.18 The sequences involved in thedelayed downregulation of cyr61 and ctgfexpression have not been identified as yet.

Although the expression of CEF10/âIG-M1(the mouse equivalent of cyr61) and âIG-M2(the mouse equivalent of ctgf) were reported tobe inducible by TGFâ in AKR-2B mouseembryo fibroblasts,7 the expression of cyr61was not found to be significantly induced byTGFâ in human skin fibroblasts or NIH3T3cells.38

These “conflicting” results indicated that theregulation of CCN gene expression may vary indiVerent species (cef10 and cyr61), anddepend upon the type of cells used in the samespecies (AKR-2B versus NIH3T3). Theseobservations stressed once more the fact thatone must be extremely cautious in comparingresults obtained with diVerent systems.

Given the importance of TGFâ signalling inseveral fundamental biological processes,much attention has been paid to the relationbetween TGFâ and ctgf expression. Promoteranalysis of ctgf and methylation interferenceassays have led to the identification of a 13nucleotide TGFâ putative responsive element(TbRE), which is conserved in the promoter ofctgf throughout species and appears to be spe-cific for TGFâ induced transcriptional activa-tion of ctgf.38 Whether binding of a specificfactor to this sequence is responsible for theprolonged induction of ctgf mRNA after shortterm exposure to TGFâ remains to beestablished.

The expression of nov was not aVected byTGFâ in either NRK-49 F (D Lawrence and BPerbal, 1995, unpublished observations) orNIH3T3 and human skin fibroblasts.38

Ex vivo experiments performed with thehuman nov promoter indicated that sequenceslocalised between positions −405 and −62 hada negative eVect on transcriptional activitywhen measured by chloramphenicol acetyltransferase (CAT) assays.11 A comparison ofthe promoter sequences from xenopus,chicken, and human nov did not reveal anysignificant conservation (fig 7). A stretch of 19nucleotides (GCAGGCCCCGGCGCGCCGC) appeared to be conserved in the humanand chicken nov promoter. Whether this novconserved element (NCE) binds a regulatoryfactor involved in the negative regulation of novpromoter activity is under investigation. Theactivity of the human nov promoter was down-regulated indirectly by the WT1 protein.39 It isworth noting that the ctgf promoter was alsorecently reported to be a target for WT1 tran-scriptional regulation.43 Screening of oligonu-cleotide arrays and northern blotting estab-lished that ctgf expression was upregulated in aWilms’s tumour cell line (WiT49A) transfectedby a dominant negative mutant of WT1,whereas ctgf promoter activity was repressedby wild-type WT1 protein. As established inthe case of nov, the downregulation of ctgf pro-moter activity was not mediated by canonicalWT1 recognition elements.43 These resultsreinforced the possibility that the alteration of

68 Perbal

www.molpath.com

on Decem


http://mp.bm

j.com/

Mol P


pril 2001. Dow

nloaded from

http://mp.bmj.com/

nov expression observed in avian nephroblasto-mas was indeed related to impaired kidneyblastema diVerentiation and tumour develop-ment.

EXPRESSION PATTERN IN NORMAL EMBRYONIC

AND ADULT TISSUES

Early studies indicated that the CCN geneswere diVerentially expressed in normal tissuesduring development.

Northern blot analysis of chicken embryonicand adult total RNA established that theexpression of the nov gene was regulated bothtemporally and spatially during develop-ment,9 12 with nov RNA detected either only atthe embryonic stage (heart) and adult stage(lung), or at both stages (brain). Muscle andintestine were also found to express nov at theembryonic stage, whereas adult spleen waspositive for nov expression. Embryonic kidneysexpressed high amounts of nov, whereas muchlower amounts were detected in the kidneys ofbirds after hatching.9 In human tissue samples,abundant amounts of polyA nov RNA weredetected by northern blotting in heart, brain,prostate, testes, and pancreas, whereas lowamounts were detected in spleen, ovary, smallintestine, colon, peripheral blood leucocytes,lung, skeletal muscle, and kidney.36 RNase pro-tection assays performed on total RNA of 8and 10 week old embryonic human tissuesindicated that nov was highly expressed in theadrenal gland and to a lesser extent in kidney,limb bud, and spinal cord, whereas heart, liver,intestine, lung, placenta, and testis were foundto be negative for nov expression (S Kocial-kowsky et al, 2001, in press).

A cDNA-PCR analysis of WISP RNAexpressed in adult and fetal human tissues16

established that WISP-2 RNAs were to befound in a rather restricted panel of tissues,including adult skeletal muscle, colon, ovary,and fetal lung, whereas WISP-1 RNAs weredetected in adult heart, kidney, lung, pancreas,ovary, spleen, and small intestine. In additionto heart, kidney, lung, and spleen, fetal skeletalmuscle and liver also contained WISP-1 RNAs.The WISP-3 RNAs were detected only in adultkidney and testis and in fetal kidney. No

expression of the WISP genes was detected inbrain.16 In situ hybridisation using wholemount mouse embryos and northern blotanalysis of adult rat and mice organs could notdetect rCOP1 RNA in brain, heart, lung,kidney, pancreas, spleen, intestine, stomach,and skeletal muscle.15 The expression ofWISP-2 in human skeletal muscle was in con-trast to the lack of expression of rCOP-1 inrodents.

Northern blot analysis of ctgf expression innormal adult tissues28 established that ctgfpolyA RNAs were found abundantly in spleen,ovary, gastrointestinal tract, prostate, heart,and testis, and at a lower level in thymus,placenta, lung, skeletal muscle, kidney, andpancreas. No expression of ctgf was detected inbrain, liver, and peripheral leucocytes.28 Thelack of ctgf expression in brain was in contrastto the western blotting identification of a40 kDa CTGF protein in rat brain homoge-nates and the immunochemical detection ofCTGF in the cerebral cortex, the white matterof spinal cord.44 The strongest positive signalfor CTGF detection was seen in astrocytes ofthe cerebral cortex, ependymal cells of the cer-ebral ventricle, tanycytes along the centralcanal of the spinal cord, and pyramidal cells inthe cortical layers III and V.

RNase protection assays performed on totalRNA from BALB/c mouse tissues establishedthat cyr61 expression is highest in lung;moderate in heart, uterus, and skeletal muscle;and low in kidney, adrenal gland, testes, brain,and ovary.45

Altogether, these results (see table 2 for asummary) illustrated the diversity of those tis-sues in which expression of the CCN genes canbe detected at diVerent stages of development,reinforcing the current view that CCN proteinsare involved in various biological processes.From table 2 it can be seen that NOV was theonly CCN protein highly expressed in thebrain. Indeed, in situ hybridisation and immu-nocytochemistry performed on developinghuman, chicken, and rat embryos confirmedthat the expression of nov was tightly associ-ated with nervous system diVerentiation and

Figure 7 Schematic structure of the human (A) and chicken (B) nov promoters. The putative binding sites were identified on the basis of their consensusnucleotide sequences.

A

–700

TG (20) SP1 AP1 NF1/CTPS1 AP2 NCE NF1/CTF AP2TATA SP1 NF1/CTF E1 SP1 E2

–581 –455 –434 –283 –222 –179 –139 –39 –29 –15 +30 +87 +211 +298

B

AP2AP2 NCE NFKE2 TATA SP1SP1SP1SP1 E1 SP1 SP1 E2

–235 –190 –171 –89–114 –70 –48 –40 +6 +28 +117 +185+136

ATG

ATG


www.molpath.com

on Decem


http://mp.bm

j.com/

Mol P


pril 2001. Dow

nloaded from

http://mp.bmj.com/

development (S Kocialkowsky et al, 2001, inpress; G Chevalier et al, 2000, unpublishedresults; W Cai et al, 2000, unpublished results;Y Cherel et al, Proceedings of the firstinternational workshop on the CCN family ofgenes, 17–19 October 2000, Saint-Malo,France).46 The role and functions of NOV inthe nervous system are under current investiga-tion.

In situ hybridisation and immunocytochem-istry performed on developing chicken em-bryos allowed us to establish the pattern of novexpression in kidney, muscle, nervous system,and skeletal development (G Chevalier et al,2000, unpublished results). In general, a goodmatch was seen between nov RNA and NOVprotein detection. The expression of nov wasdetected at embryonic days 3 and 4 (E3 andE4) in the mesonephric mesenchyme, epithe-lial vesicles, and glomeruli. At later stages (E6to E12), diVerentiated tubules were found tobe positive for nov. Metanephric epithelialvesicles, glomeruli, and ureteric bud were alsopositive for nov until E14. From E4 onwards,developing myotome and skeletal musclestained positive for nov. Nov expression wasalso detected in smooth muscle cells andcardiomyocytes. The nervous system wasfound to be strongly positive for nov. At earlystages (E3) neuroepithelium was positive atE3, and at later stages (E3–E7) neural tube alsoshowed strong staining for nov. In agreementwith our previous detection of high amounts ofNOV protein in human neuronal cells andaxons,35 high concentrations of NOV proteinwere detected in chicken neuronal cells, whichexpress low amounts of nov RNA, thereforereinforcing the possibility that in these cells, aswell as in podocytes, the NOV proteinaccumulates or is stabilised.35 Because novshares identity with the fruit fly axon repellentslit protein, the detection of nov along the

chicken and human axons suggests that itmight be involved in axonal guidance. Strongexpression of nov RNA and proteins was alsodetected at sites of chondrogenesis, where it isprobably involved in the late stages of chondro-cyte diVerentiation and osteogenesis (seebelow).

In situ hybridisation and immunocytochem-istry established that although the distributionof nov RNA and protein was widespread in firsttrimester human embryos (table 3), the majorsites of nov expression included the centralnervous system, the adrenal cortex, and diVer-entiating muscle and kidney (S Kocialkowskyet al, 2001, in press). In all cases, the NOV pro-tein was predominantly cytoplasmic, and weakstaining of the extracellular matrix (ECM) wasalso seen in some instances. In general, therewas a fairly good correlation between nov RNAand NOV protein expression, suggesting thatnov acts locally rather than distantly. In agree-ment with results obtained at later develop-mental stages,46 neurones of both the spinalcord and the brain expressed high amounts ofnov RNA and the NOV protein was clearlydetected in cell bodies and axons. The floorplate of the spinal cord, spinal nerves, and dor-sal root ganglia were identified as sites of abun-dant nov expression.

Muscle was the predominant mesodermalcomponent expressing nov. Myotubes and fus-ing myoblasts were identified as major sites ofnov expression in skeletal muscle. Smoothmuscle also expressed nov, but to a lesserextent. At the early stages, cartilage chondro-cytes did not express nov, whereas in the olderembryo, the NOV protein was demonstrated inthe perichondrium and in hypertrophicchondrocytes in the head, ribs, and lowerlimbs. In the developing urogenital tract, majordiscrepancies were noted between NOV pro-tein and nov RNA values, suggesting that theNOV protein might accumulate at these sitesfor a particular function, although we cannotexclude, as yet, that nov RNA species are shortlived in these cells. For instance, the concentra-tions of NOV protein in the podocytes of bothmesonephric and metanephric glomeruli areusually extremely high, whereas in the samecells nov RNA species are barely detectable (SKociazlkowski et al, 2001, in press; G Chevalieret al, 2000, unpublished results).35 Because theNOV protein is fairly stable in the conditioned

Table 2 Expression of CCN RNA in diVerent species

H BR PL LU LI S/M K PA SP TH PR T OV CO S/I LEU AD

c-nov E ++ ++ NT − − + + NT NT NT NT NT NT NT + NT NTc-nov A − ++ NT ++ − − NT NT + NT NT NT NT NT NT NT NTh-nov E − + − − − NT + NT NT NT NT NT NT NT − NT ++h-nov A ++ ++ − + − + + ++ + − ++ ++ + + + + NTctgf ++ − + + − + + + ++ + ++ + ++ ++ ++ − NTcyr61 ++ +/− NT +/− − ++ + − − NT NT +/− +/− NT − NT +WISP-1E + − NT + +/− +/− ++ NT NT − NT NT NT NT NT NT NTWISP-1A +/− − +/− +/− − − +/− +/− +/− − − − ++ − ++ − NTWISP-2E − − NT + − − − NT NT − NT NT NT NT NT NT NTWISP-2A − − − − − + − − − − − − + + − − NTWISP-3E − − NT − − − ++ NT NT − NT NT NT NT NT NT NTWISP-3A − − − − − − ++ − − − − + − − − − NTrCOP-1E − − NT − NT − − − − NT NT NT NT NT − NT NT

Origin of tissues: H, heart; BR: brain; PL, placenta; LU, lung; LI, liver; S/M, skeletal muscle; K, kidney; PA, pancreas; SP, spleen; TH, thymus; PR, prostate; T, tes-tis; OV, ovary; CO, colon; S/I, small intestine; LEU, leukocytes; AD, adrenal.c-nov, chicken nov RNA; h-nov human nov RNA; E, embryonic; A, adult; NT, not tested.

Table 3 Distribution of nov transcripts in 10 week human embryo

Ectoderm Endoderm Mesoderm

Brain ++ Gut endothelium ++ Adrenal cortex +++++Spinal cord (mt,fp) +++ Bronchial epithelium + Kidney tubules +++Cranial ganglia +++ Pancreatic ducts + Fusing skeletal muscle +++Auditory apparatus +++ Gut smooth muscle ++Dorsal root ganglia +++ IndiVerent gonad ++

Heart +Mesonephros +Perichondria +

Data are from S Kocialkowski et al (2001, in press).

70 Perbal

www.molpath.com

on Decem


http://mp.bm

j.com/

Mol P


pril 2001. Dow

nloaded from

http://mp.bmj.com/

medium of NOV producing cells (B Perbal,2000, unpublished results),35 the raised con-centrations of NOV detected in podocytesprobably results from its accumulation.Whether this accumulation results from itsinteraction with other proteins at this site or isphysiologically important (needed forglomeruli filtration?) remains to be established.The metanephric mesenchyme was stronglypositive for nov RNA expression, with increas-ing staining as diVerentiation proceeds in theS-shaped bodies, and a dramatic decrease onceglomeruli diVerentiation was achieved. Theseresults were of particular interest because theystrongly suggested that nov expression wastightly coupled to the diVerentiation process inthese cells. Examination of immunocytochemi-cal labelling in the heart indicated that at thesestages the endocardium was negative for nov,although it is in close contact with the myocar-dium, an abundant source of NOV protein.This observation provided evidence that theNOV protein acts where it is produced ratherthan being transported and accumulating at adistant site. Among endodermal derived tis-sues, the gut mucosa, bronchial epithelium,and pancreatic ducts were positive for nov(table 3).

The immunocytochemical localisation ofCYR61 and FISP12 proteins during mouseembryogenesis showed similarities and diVer-ences that were indicative of diVerential actionsin development.47 The localisation of CYR61,but not FISP12, in limb bud mesenchyme invivo and in vitro had suggested that CYR61might be involved in chondrogenesis.47 48 Puri-fied CYR61 protein was indeed shown to pro-mote the initial cell aggregation responsible formesenchymal condensation before chondro-genic diVerentiation and to enhance thesynthesis of cartilage specific matrix.49 Morerecently, it has been reported that in rat, cyr61is expressed in a time dependent manner inmesenchyme cells and osteoblasts from frac-ture callus,50 whereas conventional in situhybridisation did not detect cyr61 mRNA innormal bone. In the hFOB human osteoblastcell line, cyr61 was identified as a greatlyupregulated gene by various factors importantfor bone metabolism such as 1,25(OH)2-vitamin D3, tumour necrosis factor á (TNF-á), EGF, basic FGF, and interleukin 1â(IL-1â).39

DiVerential display PCR identified ctgf asbeing preferentially expressed in HGS-2/8human chondrosarcoma derived chondrocyticcells.51 In situ hybridisation established thatctgf was specifically expressed in the hyper-trophic chondrocytes of costal cartilage andvertebral column in 17 day old mouseembryos,51 at a time when cyr61 expression isstrongly downregulated,48 suggesting thatCYR61 and CTGF might have complemen-tary but diVerent functions. Staining of thecost-chondral junctions of newborn mouse ribsindicated that ctgf was expressed in chondro-cytes of the hypertrophic zone, but not in theproliferating and resting zone.52 Osteoblasts didnot express ctgf. Because ctgf mRNA expres-sion was stimulated in HCS2/8 cells treated

with TGFâ and bone morphogenic protein 2(BMP-2), two factors believed to be involved incartilage and bone formation, it was proposedthat CTGF was a mediator of these cytokinesin cartilage. Because CTGF promoted theproliferation and diVerentiation of chondro-cytes in culture,53 it was proposed that ctgfexpression is involved in endochondral ossifica-tion by acting on proliferating and maturatingchondrocytes in growth cartilage. Bothchondrocytic and osteoblastic cells express a280 kDa protein that has been described as areceptor for CTGF.54 55

Because nov expression was also detected insites of chondrogenesis during chicken devel-opment (fig 8), we postulated that the negativeregulatory activity attributed to nov (see above)might be involved at later stages of diVerentia-tion to counterbalance the stimulatory eVectsof CYR61 protein on chondrogenesis, whichhad been established in micromass cell cul-tures.49 According to this view, CYR61 andNOV protein activities might represent the“yin and the yang” along common regulatorypathways leading to chondrocyte diVerentia-tion and osteogenesis.

To assess the potential role of nov inchondrogenesis and osteogenesis, we first ana-lysed the expression of nov chicken duringwing and leg development.56 Northern blottingperformed on total RNA prepared from limbbuds dissected between E4 and E18 estab-lished that expression of nov increasing signifi-cantly between E8 and E10 and progressivelydecreased until E14, to reach a basal value atE15.56 The expression profile of nov correlatedwith the time course of limb development,which is completed after 10 days of incubationunder normal conditions. Sustained expressionindicated that nov was required over anextended period of time during limb develop-ment, therefore suggesting that it might beneeded for later events. Immunocytochemistryperformed on paraYn wax embedded wingbud sections established that nov was mainlydetected at E6 in the proliferating chondro-cytes and the perichondrium (fig 9A), whereasat E11/E12, staining of nov was essentiallydetected in the hypertrophic chondrocytes ofthe central diaphyseal region and to a lowerextent in the perichondrium (fig 9B). At E14,nov labelling was greatly reduced and restrictedto a few hypertrophic chondrocytes in thedeveloping wing. Flattened chondrocytes werenot positive for nov expression at any stage ofdevelopment (fig 9C). Pretreatment of the sec-tions with hyaluronidase resulted in additionalwidespread labelling of the cartilaginous ma-trix, in agreement with nov being associatedwith the ECM.57 Similarly, CYR61 and CTGFwere also reported to be ECM associated pro-teins and suggested to mediate cell–matrixinteractions.58

When mesenchymal cells isolated from thelimb buds of 3.5 day old chicken embryos orfrom 10.5 day old mouse embryos are seededat high density (107 cells/ml), cellular aggrega-tion proceeds and internodular condensationsundergo chondrocytic diVerentiation, as shownby the formation cartilaginous nodules within


www.molpath.com

on Decem


http://mp.bm

j.com/

Mol P


pril 2001. Dow

nloaded from

http://mp.bmj.com/

three days of culture. When cells are seeded ata lower density (106/ml), chondroblastic diVer-entiation does not proceed and no cartilagi-nous formation occurs. The use of this systemrevealed that CYR61 enhances chondrogenicdiVerentiation.49 Northern blotting experi-ments performed on chicken micromass cell

cultures established that nov expression, whichwas not detected at day 0 of culture, increasedprogressively for two days and decreased at day3, whereas the expression of collagen II, amarker for determination of progenitor cells toundergo chondroblastic diVerentiation, in-creases from day 1 onwards (fig 10). The

Figure 8 Detection of nov RNA species at sites of chondrogenesis in developing chicken. (A) Development stage (St) 3,embryonic day (E) 8.5. (B) St 31, E7. (C) St 35, E8.5. (D) St 38, E12. (E) St 35, E8.5.

72 Perbal

www.molpath.com

on Decem


http://mp.bm

j.com/

Mol P


pril 2001. Dow

nloaded from

http://mp.bmj.com/

expression of nov is dependent upon theproductive diVerentiation of mesenchymecells, as shown by the lack of nov expression inthe absence of cartilaginous nodule formationat low cell density (fig 11). When cells wereseeded at high density, ctgf expression wasdetected in micromass cultures from day 2onwards, whereas CEF10 expression wasdetected from day 0 to day 3.49

These observations and the results of immu-nochemical studies performed on sections sug-gested that NOV and CTGF were required atlate stages of the chondrogenesis/osteogenesisdiVerentiation process, whereas CYR61 wasrequired at early stages. Whether the diVeren-tial expression of these three CCN genes in theskeletal system is temporally related remains tobe established.

It is worth noting that in addition tonephroblastoma, MAV can induce osteopetro-sis in chickens, an abnormal proliferation ofosteoblasts leading to severe bone diseases.10

Based on the consideration that nov was a tar-get of MAV in chicken blastemal cells, it istempting to hypothesise that the induction ofosteopetrosis by MAV could also result from analteration of nov expression in bone precursorcells.

EXPRESSION PATTERN IN PATHOLOGICAL

CONDITIONS

Evidence indicating that CCN proteins partici-pate in various diseases has been accumulatingat an increasing pace over the past years.

The relation that was established betweenctgf expression and TGFâ has resulted in con-siderable interest in the potential role of CTGFin fibrosis and its possible medical applications.This has led to the widely accepted conclusionthat the profibrogenic properties of TGFâresult from ctgf induction, stimulation of fibro-blast proliferation, and overproduction ofextracellular matrix. As discussed in depth byEssam et al,4 the relation between ctgf andTGFâ expression is not absolute and there are

Figure 10 Expression of nov RNA species and collagen II (COL II) in chickenmicromass cell cultures. Staining of ribosomal RNA species allows relative quantification ofthe hybridisation signals obtained in two independent experiments.56 J0–J3, days of culture;GAPDH, glyceraldehyde-3-phosphate dehydrogenase.

Nov 2.2 kb

4.2 kb

1.3 kb

28S

18S

COL II

GAPDH

Et. Br.

C F J0 J1 J3J2 J2J0 J1 J3

Figure 9 Detection of the NOV protein in developing bone. (A) At stage 29 (embryonic day 6 (E6)), NOV is detected atsites of chondrogenesis in the developing chicken wing. Alcian blue staining identifies cartilage formation. (B) At stages35–38 (E11–12) strong nov staining is seen in the perichondrium (p) and proliferating chondrocytes (pc). (C) At the samestages, flattened chondrocytes (fc) do not stain positive for nov, whereas hypertrophic chondrocytes (hc) are strongly positivefor NOV staining (D).


www.molpath.com

on Decem


http://mp.bm

j.com/

Mol P


pril 2001. Dow

nloaded from

http://mp.bmj.com/

several studies indicating that ctgf expressionmay be independent of TGFâ in fibroblasts.

Nonetheless, ctgf expression is involved inmany types of diseases, including fibroticdiseases, atherosclerosis, inflammatory dis-eases, and scleroderma. The expression of ctgfalso increases in wound healing. Because theseaspects of CTGF biology have been reviewedextensively,2–4 I will focus on studies that iden-tify CCN proteins as key players in tumourdevelopment.

For many years, the only evidence suggestingthat CCN genes might be involved in cancerdevelopment came from studies performedwith nov.

Because nov was originally shown to behighly expressed in all avian nephroblastomas,we initiated studies aimed at establishingwhether nov expression was also altered inhuman Wilms’s tumours and whether analteration of nov expression was associatedwith other types of tumours. These studieswere conducted with two hopes, namely: (1)that measurement of nov in biological fluids orin tumours might be of clinical value for earlydiagnosis, typing, and prognosis; and (2) thatthe antiproliferative activity of nov could beused in cancer treatment.

The studies that we have performed withWilms’s tumours established that they con-tained abnormal amounts of nov RNA, whichwere in some cases inversely related to that ofthe WT1 RNAs detected in the same sam-ples,59 an observation that, at first glance, was

in agreement with the ex vivo downregulationof nov promoter activity by WT1.39 However,studies performed with a larger panel ofsamples35 representative of sporadic, Wilms,aniridia, growth retardation (WAGR) andDenis Drash syndrome (DDS) histologicaltypes of tumours did not confirm that inverserelation and raised the possibility that the vari-ations of nov expression resulted from diVerentrelative amounts of WT1 isoforms contained inthese tumours.35

These studies also established that novexpression was a marker of heterotypic diVer-entiation.35 In striated muscle tumours, theexpression of nov was shown to occur at anearlier stage than desmin during the hetero-typic diVerentiation of the blastemal cells thattakes place in these tumours,35 and confocalmicroscopy established that the NOV proteinwas colocalised with desmin in heterotypicmuscle fibres (G Chevalier, 1998, unpublishedobservation).

Studies that we have performed with severalother types of tumours have established astrong association between nov expression andtumour diVerentiation in rhabdomyosarcomas,neuroblastomas, and cartilage tumours (HYeger et al, and C Yu et al, Proceedings of thefirst international workshop on the CCN fam-ily of genes, 17–19 October 2000, Saint-Malo,France). In the case of neuroblastomas withpoor prognostic features, nov staining was low/moderate within the small tumour cells,whereas in tumours with a favourable progno-sis and no N-myc amplification, nov stainingwas strong in the cytoplasm of diVerentiatedganglion-type cells (H Yeger et al, 1998,unpublished results). These results indicatedthat nov expression can be used as a valuablemolecular marker for neuroblastoma typingand for tumour prognosis.

Because the adrenal was identified as a majorsite of nov expression in the human embryo(see above), and because there is no molecularprognosis factor for adrenocortical carcinoma,we examined whether alterations in novexpression occurred in these highly lethal can-cers with a mortality rate over 50%. In the nor-mal adult adrenal cortex, immunocytochemi-cal studies showed a distinct positive labellingfor nov in the zona fasciculata, the middle layerof the adrenal cortex. The labelling was fairlyuniform and confined to the cytoplasm (fig12A). In the adrenocortical adenomas therewas clear staining of nov in cells that probablymimic the zona fasciculata (fig 12B). However,staining in the adenomas was not homogenous,therefore suggesting that the amount of novdetected in the cells of these benign tumourswas dependent upon their metabolic state orwas related to hormone production. Adreno-cortical carcinomas stained either negative (fig12C) or positive for nov. No clearcut relationcould be established between the degree of novexpression and tumour stage.

In pronounced contrast and apparent con-tradiction to these reports, our studies alsoestablished that the expression of nov corre-lated with the metastatic potential of tumourcells in the case of Ewing’s sarcomas (MC

Figure 11 The expression of nov RNA depends uponchondrogenesis. (A) 106 mesenchymal cells were seeded formicromass cell cultures. (B) 107 mesenchymal cells wereseeded for micromass cell cultures. Samples were taken atthe day of seeding (J0) and after 24 hours culture (J1).COL II, collagen II; GAPDH, glyceraldehyde-3-phosphatedehydrogenase.

COL II

NOV

GAPDH

J0 J1

a b

J0 J1

74 Perbal

www.molpath.com

on Decem


http://mp.bm

j.com/

Mol P


pril 2001. Dow

nloaded from

http://mp.bmj.com/

Manara et al, 2000, unpublished results), pros-tate cancer (B Cadot et al, Proceedings of thefirst international workshop on the CCN fam-ily of genes, 17–19 October 2000, submittedfor publication) and renal cell carcinoma (LGlukhova et al, 2001, submitted for publi-cation; L Glukhova et al, Proceedings of thefirst international workshop on the CCN fam-ily of genes, 17–19 October 2000, submittedfor publication). In addition to the labelling ofNOV in the cytoplasm of acinar epithelial cells,prostate hyperplasia showed intense luminallabelling suggestive of NOV secretion in semi-nal fluid. Whether measurement of NOV inseminal fluid can be of value as a marker ofprostate disease is currently being investigated.In prostate tumour cell lines, the expression ofnov was detected in the cytoplasm of three celllines derived from metastasis to bone (PC3),brain (DU145), and lymph node (LNCap). Noexpression was detected in the SV40-T antigenimmortalised PNT1B cells, an observation incontrast to a recent study indicating thatSV40-T immortalised P69 cells were positivefor nov expression.60 In renal cell carcinomas(RCCs), which represent 85–90% of all kidneytumours, a significantly higher concentration

of secreted NOV protein was detected by west-ern blotting in the conditioned medium of fastgrowing tumours. These results indicated thatthere was an inverse relation between theamount of NOV secreted by the tumour cellsand the time that they required to establishtumours in vivo and for the tumour to reach agiven size (fig 13).

Finally, results that were obtained in ourstudy of glioblastoma cell lines suggested thatnov expression was correlated with reducedtumorigencity and reduced metastatic poten-tial in these brain tumour cells.61 Whether wecan use the negative growth regulatory proper-ties attributed to nov as a tool for treatment ofthese highly aggressive tumours is currently thematter of intense investigation.

Therefore, it appeared that the expression ofnov was altered in many types of tumours andthat high concentrations of NOV could beassociated with either diVerentiation or in-creased proliferation and metastasis. Thiscomplex and apparently conflicting situation isnot unique to nov among the CCN proteins. Asreported above: (1) the WISP genes whoseexpression is upregulated in Wnt transformedcells are either upregulated or downregulatedin colon cancers, and (2) the expression ofWISP-1 increased the tumorogenicity of trans-fected cells whereas overexpression of Elm1had opposite eVects.

It is widely accepted that the establishmentand development of tumours rely on the abilityof the tumour cells to overcome hypoxia andestablish a network of blood vessels that willultimately permit the tumour to expand. Neo-vascularisation is therefore controlled by thetumour cells that secrete angiogenic factors toattract endothelial cells, which in turn arebelieved to produce growth factors stimulatingtumour growth. Several positive and negativeregulators of angiogenesis have been described.The angiogenic properties of CYR61 andCTGF2 62 63 suggest that these proteins mightbe involved in tumour growth. Indeed, gastricadenocarcinoma RF-1 cells transfected withcyr61 gave rise to larger and more vascularisedtumours than parental cells when injected intonude mice.62 Many tumour cell lines werereported to express cyr61 at various levels.Although it was reported that cyr61 expressingtumour cells were more tumorigenic than those

Figure 12 Detection of the NOV protein in normal (A)and tumoral adrenal tissues (B and C). The normal areawas isolated from a patient with adenoma. Note the unevennov staining in the adenoma (B). An example of NOVnegative adrenal carcinoma is shown (C). Photographswere kindly provided by H Yeger.

Figure 13 Expression of nov in renal cell carcinomas(RCCs). Increased amounts of nov were correlated withincreased tumorigenicity. Nov, levels of nov RNA estimatedfrom northern blots; Latency, time in days required to reacha tumour of a given size.

200

150

100

50

0

250

200

150

100

50

0RCC

7

Latency

Passage

Nov

RCC44

RCC47

RCC1

RCC49

RCC41

Day

s No

v


www.molpath.com

on Decem


http://mp.bm

j.com/

Mol P


pril 2001. Dow

nloaded from

http://mp.bmj.com/

that did not express this gene,2 no clearcutassociation of this kind was drawn from studiesthat we performed with glioblastoma celllines.64 Downregulation of cyr61 expressionwas reported in rhabdomyosarcomas,65 and inabout 50% of prostate tumours samples,66

whereas increase expression of cyr61 has beendetected in pancreatic cancers (C Wenger andTM Gress, Proceedings of the first inter-national workshop on the CCN family ofgenes, 17–19 October 2000, Saint-Malo,France) and in about 30% of invasive breastcarcinomas.67

Increased expression of ctgf has also beendetected by northern blotting in about 70% of11 human invasive mammary ductal carcino-mas,68 in 100% of nine dermatofibromasexamined, in pyogenic granuloma, in endothe-lial cells of angiolipomas and angioleiomyo-mas,69 and in 15 of 19 pancreatic tumours.70

However, five of seven cases of dermatofibro-sarcoma protuberans and two cases of malig-nant fibrous histiocytoma were negative for ctgfexpression.69 A more recent immunocyto-chemical study performed with 18 chondro-sarcomas representative of various histologicalgrades established that ctgf expression closelycorrelated with increasing malignancy.71

Although the number of reports concerningthe expression of cyr61 and ctgf in tumours isstill limited, the general picture is that neither astraightforward nor a general correlation canbe drawn between the expression of cyr61 orctgf and tumour development.

A clue to understanding the discrepanciesobserved for the expression of the CCN genesin cancer cells might come from the multi-modular structure of the CCN proteins andtheir potential ability to interact sequentially orsimultaneously with diVerent partners, thebioavailability of which might be a key elementin the manifestation of the positive or negativeeVects attributed to the CCN proteins. In line

with this, I would like to propose that the bio-logical activity of the CCN proteins is depend-ent upon diVerent combinatorial eVects.Therefore, a search for proteins and ligandsinteracting with nov has been undertaken.

Identification of CCN protein partnersAs a first step in identifying the proteins thatmight interact with NOV and other CCN pro-teins we have used30 33 the two hybrid systemstrategy, which has been used successfully overthe past few years to uncover many proteininteractions involved in cell growth signalling.

To avoid selecting targets that might onlyinteract with isolated modules, we used the fulllength NOV protein fused to the DNA bindingdomain of Gal4 DNA as bait. The interactionsthat were detected with this system wereusually validated by pull down assays per-formed with a panel of glutathioneS-transferase (GST) fusion proteins containingdiVerent portions of nov. The use of theserecombinant proteins also allowed us toidentify the domains of NOV that wereinvolved in the interactions.

In addition to RNA polymerase II subunit7,30 we identified fibulin 1C as a protein inter-acting with NOV.33 The potential interaction ofNOV and fibulin 1C provided the first evidencefor a role of NOV in cell adhesion signalling.Fibulin 1C is a 100 kDa protein of the ECMthat has been reported to bind fibronectin,laminin, fibrinogen, and basement membraneprotein nidogen.33 The expression of fibulin 1in the perichondrium and calcifying regions ofdeveloping bones, in the gut subepithelium,and in peripheral nerves of human embryos ofgestational weeks 4–10 correlated with sites ofnov expression at the same developmentalstages, and provided another indication thatNOV and fibulin might indeed interact in vivo.Furthermore, the levels of nov and fibulin 1CRNAs correlated inversely in several cell lines,33

and ectopic expression of nov is accompaniedby a decrease of fibulin 1C concentrations,therefore suggesting that a downregulation offibulin 1C transcription might take place incells expressing high amounts of nov.

These studies also established that theC-terminal module of NOV was suYcient topromote the interaction of NOV with fibulin1C. The association of NOV and fibulinthrough the C-terminal portion of fibulin 1C,which contains five of nine EGF repeatspresent in fibulin 1C, raised the possibility thatthe binding of NOV might mediate the poten-tial role of fibulin 1C in the assembly of theECM and cell adhesion.33 It is worth notingthat the C-terminal domain of NOV was alsoshown to mediate the interaction of NOV withCTGF in the two hybrid system, thereforesuggesting that heterodimerisation of CCNproteins might take place in vivo and that crosstalk between CCN proteins is involved in themodulation of their biological activities. CCNproteins such as rCOP1 and CTGF-L,72 whichare deprived of the C-terminal domain, mightplay an important role in these processes andexhibit a whole range of dominant negativetype activities.

Figure 14 Model for the structural organisation of the CCN proteins. Potentialinteractions with positive and negative regulators are thought to be disrupted upontruncation of the N-proximal and C-proximal modules of CCN proteins. Ct, cystein knotcontaining family of growth regulators-like domain; IGFBP, insulin-like growth factorbinding protein-like domain; TSP1, thrombospondin-like domain; and VWC, VonWillebrand factor-like domain.

VWC

Hinge

Hinge

Ct

Negative factors?

?

TSP1

TSP1

TSP1

VWC

Hinge

Positive factors

IGFBP

IGFBP

Ct

76 Perbal

www.molpath.com

on Decem


http://mp.bm

j.com/

Mol P


pril 2001. Dow

nloaded from

http://mp.bmj.com/

Our working hypothesis is based on theassumption that under native conditions theCCN proteins adopt a compact structure (fig14), where two halves interact with each otheras a result of the hinge region that is foundbetween modules 3 and 4 (sharing identity withvon Willebrand factor type C repeat andthrombospondin, respectively). This interac-tion would permit the N-proximal half to exerta negative regulatory control on the stimula-tory properties of the C-proximal half. Thefolded structure would also allow a certainnumber of interactions with positive and nega-tive eVectors. The bioavailability of these eVec-tors, coupled with the temporal and spatialexpression of the CCN proteins, would be cru-cial factors that dictate the constitution of thecomplexes that can be formed, and subse-quently the nature of the signals transmitted bythese complexes. Truncations of the CCN pro-teins at either their N-terminus or C-terminuswould disrupt these regulatory interactions,alter the balance of positive and negative regu-latory signals (fig 15), and result in conferringto these proteins a constitutive repressive orstimulatory eVect on cell growth, as exempli-fied by the case of WISP-2/rCOP1, and by theproduction of an oncogenic truncated NOVprotein in MAV induced nephroblastoma (seeabove). The implications of this working modelare presently being tested.

Future directions and prospectsIn the light of the recent results reported in thisreview, it seems that CCN proteins might act asmultipotent matchmakers, allowing the scaf-folding of a complex network of regulatoryproteins responsible for the regulation ofseveral fundamental functions during normallife, development, and probably death.

The identification of other proteins poten-tially interacting with NOV has established thateach of the four constitutive domains of thenative protein is required alone or in combina-tion with others to promote the interactions. Insome cases, only the full length NOV proteinwas found to interact with potential targets.Deciphering this complex situation will cer-tainly lead to the identification of major players

responsible for the regulation of NOV biologi-cal activity. Establishing whether other CCNproteins also interact with the NOV targetsremains a major goal.

More generally, there is a dramatic need forcomparative studies to be performed withdiVerent members of the CCN family. Theavailability of the reagents and the develop-ment of biological assays for several membersof the CCN family should greatly ease suchstudies, the potential value of which is obviousin the light of the diversity of the systems usedthus far.

To answer the question “what is nov doing?”I have embarked my laboratory on a pathaimed at examining several complementaryaspects of nov biology. Similar approaches per-formed with other CCN genes would certainlyprove to be extremely fruitful. Thus, I think it isessential to establish where, how, and withwhom the CCN proteins act to answer later thequestions regarding the regulation of theiractivity and the consequences of alteredexpression.

WHERE?The use of diVerent eYcient expressionsystems (such as baculoviruses) enables theproduction of recombinant tagged proteinsthat can be used to identify the sites of CCNprotein binding on cells and follow the fate ofthe secreted proteins once they are presented atthe cell surface. The production of specificpolyclonal and monoclonal antibodies directedagainst diVerent portions of the CCN proteinsalso enables their concentrations to bemeasured in biological fluids. They can alsoestablish whether quantitative variations of cir-culating CCN proteins are associated withpathological conditions and can be used as avaluable clinical tool. We are presently usingthis strategy to measure, under pathologicalconditions, the variations of NOV released indiVerent biological fluids, such as cerebrospi-nal fluid, seminal fluid, urine, and serum.

HOW?This aspect requires the purification of nativeCCN proteins from a reliable biologicalsource. Although the use of recombinantproteins has proved very fruitful in unravellingthe biological properties of several signallingeVectors, it must be kept in mind that manybiochemical properties of native proteins re-quire post translational modifications, whichdo not take place (or only partially) in most ofthe systems used to produce recombinant pro-teins. Purification of the native protein alsopermits the isolation of functional complexeswhose characterisation has proved to beextremely informative in many instances.Because CCN proteins potentially interactwith several other partners, the isolation ofcomplexes in which these proteins are engagedwill undoubtely lead to important discoveriesconcerning their mode of action. The purifica-tion of the diVerent CCN proteins can alsoestablish whether they exhibit synergistic orantagonistic properties when tested in combi-nation.

Figure 15 CCN proteins and cancer. This schematic drawing summarises the diversity ofsituations encountered for CCN proteins in tumours. Whether the subcellular localisation ofthe CCN proteins other than NOV can result in opposing biological properties remains to beestablished.

Upregulated, suppressornov

Unregulated, inducderwisp1

Downregulated, suppressorElm1

Upregulatedoncogenic

nov

Upregulatedwisp2

Downregulated, suppressorWisp2/rCOP-1

Full length

NH2 truncated

COOH truncated


www.molpath.com

on Decem


http://mp.bm

j.com/

Mol P


pril 2001. Dow

nloaded from

http://mp.bmj.com/

WITH WHOM?

Immunoprecipitation of complexes, covalentcrosslinking, and solid phase assays are amongthe biochemical strategies being used currentlyto identify ligands interacting with the CCNproteins. The use of biological systems such asphage display, the two hybrid system, etcshould also permit the identification of proteinsinteracting with some, or many members, ofthe CCN family. Establishing whether crossinteractions exist among the diVerent CCNproteins will be of great interest.

Questions regarding the biological activity ofthe CCN proteins in vitro or ex vivo are alsoworth considering. There are several systems inwhich to study the eVects of CCN proteins onproliferation, diVerentiation, and cell death.Again, comparative studies performed withdiVerent members of the CCN family in thesame system should greatly improve ourunderstanding of the biology of CCN proteins.

Post transcriptional and post translationalprocesses are thought to be involved in theregulation of the bioavailability and functionsof CCN proteins. PKC is involved in the regu-lation of CCN RNA post translational turno-ver. Activation of PKC has recently beenreported to inhibit the expression of ctgf.73 Itwould be worth examining more precisely therelation that might exist between PKC directedphosphorylation and CCN gene expression.

Last, but not least, the bulk of studies aimedat establishing whether the CCN proteins areinvolved, directly or indirectly, in tumourdevelopment have demonstrated the extremecomplexity and diversity of CCN gene expres-sion patterns in the various systems analysed.Again, comparative studies would certainly bevery informative and would probably allow usto establish a solid basis for the understandingof the biological properties of the CCNproteins and to determine whether they can beused as tools for future molecular medicine.

I wish to thank particularly my wife Annick for her help andconstant support. Thanks are also due to S Gabaron for readingthe manuscript. The research performed in my laboratory wasfunded by grants from the Ligue Nationale Contre le Cancer(Comités National, du Cher et de l’Indre), AssociationFrançaise contre les Myopathies (AFM), Matra-Hachettte,Fondation pour la Recherche Médicale (FRM), and Associationpour la Recherche contre le Cancer (ARC). I am grateful toProfessor B Roizman for his support and for providing the facil-ity to perform experiments with herpes simplex virus and thetwo hybrid system.

1 Bork P. The modular architecture of a new family of growthregulators related to connective tissue growth factor. FEBSLett 1993;327:125–30.

2 Lau LF, Lam SC. The CCN family of angiogenicregulators: the connection. Exp Cell Res 1999; 248:44–57.

3 Brigstock DR. The connective tissue growth factor/cysteine-rich 61/nephroblastoma overexpressed (CCN) family.Endocr Rev 1999;20:189–206.

4 Essam ED, Moussad A, Brigstock D. Connective tissuegrowth factor: what’s in a name. Mol Genet Metab 2000;71:276–92.

5 Lau LF, Nathans D. Expression of a set of growth-relatedimmediate early genes in BALB/c 3T3 cells: coordinateregulation with c-fos or c-myc. Proc Natl Acad Sci USA1987;84:1182–6.

6 Almendral J, Sommer D, Mac Donald-Bravo H, et al. Com-plexity of the early genetic response to growth factors inmouse fibroblasts. Mol Cell Biol 1998;8:2140–8.

7 Brunner A, Chinn C, Neubauer M, et al. Identification of agene family regulated by transforming growth factor-â.DNA Cell Biol 1991;10:293–300.

8 Matsunaga E. Genetics of Wilms’ tumor. Hum Genet 1981;57:231–46.

9 Perbal B. Contribution of MAV-1-induced nephroblastomato the study of genes involved in human Wilms’ tumordevelopment. Crit Rev Oncog 1994;5:589–613.

10 Perbal B. Pathogenic potential of myeloblastosis associatedviruses. Infect Agents Dis 1995;4:212–27.

11 Chevalier G, Perbal B. Genetic alterations associated withpathologic diVerentiation of Wilms’ tumors. Bull Cancer(Paris) 1997;84:289–303.

12 Joliot V, Martinerie C, Dambrine G, et al. Proviralrearrangements and overexpression of a new cellular gene(nov) in myeloblastosis-associated virus type 1-inducednephroblastomas. Mol Cell Biol 1992;12:10–21.

13 Martinerie C, Viegas-Péquignot E, Guénard I, et al. Physicalmapping of human loci homologous to the chicken novprotooncogene. Oncogene 1992;7:2529–34.

14 Hashimoto BY, Shindo-Okada N, Tani M, et al. Expressionof the ELM1 gene, a novel gene of the CCN (connectivetissue growth factor, cyr61/cef10, and neuroblastoma over-expressed gene) family, suppresses in vivo tumor growthand metastasis of K-1735 murine melanoma cells. J ExpMed 1998;187:289–96.

15 Zhang R, Averboukh L, Zhu W, et al. Identification ofrCop-1, a new member of the CCN protein family, as anegative regulator for cell transformation. Mol Cell Biol1998;18:6131–41.

16 Pennica D, Swanson TA, Welsh JW, et al. WISP genes aremembers of the connective tissue growth factor family thatare up-regulated in Wnt-1-transformed cells and aber-rantly expressed in human colon tumors. Proc Natl Acad SciU S A 1998;95:14717–22.

17 Simmons D, Levy D, Yannoni Y, et al. Identification of aphorbol ester-repressible v-w-c- inducible gene. Proc NatlAcad Sci USA 1989;86:1178–82.

18 Scholz G, Martinerie C, Perbal B, et al. Transcriptionaldown regulation of the nov proto-oncogene in fibroblaststransformed by p60v-src. Mol Cell Biol 1996;16:481–6.

19 Bradham DM, Igarashi A, Potter RL, et al. Connectivetissue growth factor: a cysteine-rich mitogen secreted byhuman vascular endothelial cells is related to the SRC-induced immediate early gene product CEF-10. J Cell Biol1991;114:1285–94.

20 Ryseck RP, Macdonald-Bravo H, Mattei MG, et al.Structure, mapping, and expression of fisp-12, a growthfactor-inducible gene encoding a secreted cysteine-richprotein. Cell Growth DiVer 1991;2:225–33.

21 Lin J, Liliensiek B, Kanitz M, et al. Molecular cloning ofgenes diVerentially regulated by TNF-á in bovine aorticendothelial cells, fibroblasts and smooth muscle cells. Car-diovasc Res 1998;38:802–3.

22 SteVen CL, Ball-Mirth DK, Harding PA, et al. Characteri-sation of cell-associated and soluble forms of connectivetissue growth factor (CTGF) produced by fibroblast cellsin vitro. Growth Factors 1998;15:199–213.

23 Igarashi A, Okochi H, Bradham DM, et al. Regulation ofconnective tissue growth factor gene expression in humanskin fibroblasts and during wound repair. Mol Cell Biol1993;4:637–45.

24 Hishikawa K, Oemar BS, Tanner FC, et al. Overexpressionof connective tissue growth factor gene induces apoptosisin human aortic smooth muscle cells. Circulation 1999;100:2108–112.

25 Hishikawa K, Oemar BS, Tanner FC, et al. Connectivetissue growth factor induces apoptosis in human breastcancer cell line MCF-7. J Biol Chem 1999;274:37461–6.

26 Hishikawa K, Nakaki T, Fujii T. Connective tissue growthfactor induces apoptosis via caspase 3 in cultured humanaortic smooth muscle cells. Eur J Pharmacol 2000;392:19–22.

27 Hurvitz JR, Suwairi WM, Van Hul W, et al. Mutations in theCCN gene family member WISP3 cause progressive pseu-dorheumatoid dysplasia. Nat Genet 1999;23:94–8.

28 Kim HS, Nagalla SR, Oh Y, et al. Identification of a familyof low-aYnity insulin-like growth factor binding proteins(IGFBPs): characterisation of connective tissue growthfactor as a member of the IGFBP super family. Proc NatlAcad Sci U S A 1997;94:12981–6.

29 Grotendorst G, Lau L, Perbal B. CCN proteins are distinctfrom and should not be considered members of theinsulin-like growth factor-binding protein superfamily.Endocrinologie 2000;141:2254–6.

30 Perbal B. Nuclear localization of NOV protein: a potentialrole for nov in the regulation of gene expression. J ClinPathol: Mol Pathol 1999;52:84–91.

31 Roizman B, Sears A. Herpes simplex viruses and their repli-cation. In: Fields BN, Knipe DM, eds. Fields virology, Vol.2. New York: Raven Press, 1990:1795–841.

32 Mahony D, Kalionis B, Antalis TM. Plasminogen activatorinhibitor type-2 (PAI-2) gene transcription requires a novelNF-KappaB-like transcriptional regulatory motif. Eur JBiochem 1999;263:765–72.

33 Perbal B, Martinerie C, Sainson R, et al. The C-terminaldomain of the regulatory protein NOVH is suYcient topromote interaction with fibulin 1C: a clue for a role ofNOVH in cell-adhesion signaling. Proc Natl Acad Sci U S A1999;96:869–74.

34 Ball DK, Surveyor GA, Diehl JR, et al. Characterisation of16-to 20-kilodalton (kDa) connective tissue growth factors(CTGFs) and demonstration of proteolytic activity for38-kDa CTGF in pig uterine luminal flushings. Biol Reprod1998;59:828–35.

35 Chevalier G, Yeger H, Martinerie C, et al. NovH: diVerentialexpression in developing kidney and Wilms’ tumors. Am JPathol 1998;152:1563–75.

78 Perbal

www.molpath.com

on Decem


http://mp.bm

j.com/

Mol P


pril 2001. Dow

nloaded from

http://mp.bmj.com/

36 Burren C, Wilson E, Hwa V, et al. Binding properties anddistribution of insulin-like growth factor binding protein-related protein 3 (IGFBP-rP3/NovH), an additional mem-ber of the IGFBP superfamily. J Clin Endocrinol Metab1999;84:1096–103.

37 Latinkic B, O’Brien P, Lau L. Promoter function and struc-ture of the growth factor-inducible immediate early genecyr61. Nucleic Acids Res 1991;19:3261–7.

38 Grotendorst G, Okochi H, Hayashi N. A novel transforminggrowth factor â response element controls the expression ofthe connective tissue growth factor gene. Cell Growth DiVer1996;7:469–80.

39 Martinerie C, Chevalier G , Rauscher F, Jr, et al. Regulationof nov by WT1: a potential role for nov in nephrogenesis.Oncogene 1996;12:1479–92.

40 Schutze N, Lechner A, Groll C, et al. The human analog ofmurine cysteine-rich protein 61 is a 1á,25-dihydroxyvitamin D3 responsive immediate early gene inhuman fetal osteoblasts: regulation by cytokines, growthfactors, and serum. Endocrinology 1998;139:1761–70.

41 Rivera-Gonzalez R, Petersen DN, Tkalcevic G, et al.Estrogen-induced genes in the uterus of ovariectomizedrats and their regulation by droloxifene and tamoxifen. JSteroid Biochem Mol Biol 1998;64:13–24.

42 Chung KC, Ahn YS. Expression of immediate early genecyr61 during the diVerentiation of immortalized embryonichippocampal neuronal cells. Neurosci Lett 1998;255:155–8.

43 Stanhope-Baker P, Williams BR. Identification of connec-tive tissue growth factor as a target of WT1 transcriptionalregulation. J Biol Chem 2000;275:38139–50.

44 Kondo Y, Nakanishi T, Takigawa M, et al. Immuno-histochemical localization of connective tissue growthfactor in the rat central nervous system. Brain Res 1999;834:146–51.

45 O’Brien TP, Yang GP, Sanders L, et al. Expression of cyr61,a growth factor-inducible immediate–early gene. Mol CellBiol 1990;10:3569–77.

46 Su BY, Cai WQ, Zhang CG, et al. A developmental study ofnovH gene expression in human central nervous system. CR Acad Sci Paris 1998;321:883–92.

47 Kireeva M, Latinkic B, Kolesnikova T, et al. Cyr61 andFisp12 are both ECM-associated signaling molecules:activities, metabolism, and localization during develop-ment. Exp Cell Res 1997;233:63–77.

48 O’Brien TP, Lau LF. Expression of the growth factor-inducible immediate early gene cyr61 correlates with chon-drogenesis during mouse embryonic development. CellGrowth DiVer 1992;3:645–54.

49 Wong M, Kireeva M, Kolesnikova T, et al. Cyr61, product ofa growth factor-inducible immediate–early gene, regulateschondrogenesis in mouse limb bud mesenchymal cells. DevBiol 1997;192:492–508.

50 Hadjiargyrou M, Ahrens W, Rubin CT. Temporal expres-sion of the chondrogenic and angiogenic growth factorCYR61 during fracture repair. J Bone Miner Res 2000;15:1014–23.

51 Nakanishi T, Kimura Y, Tamura T, et al. Cloning of amRNA preferentially expressed in chondrocytes by diVer-ential display-PCR from a human chondrocytic cell linethat is identical with connective tissue growth factor(CTGF) mRNA. Biochem Biophys Res Commun 1997;234:206–10.

52 Shimo T, Nakanishi T, Kimura Y, et al. Inhibition of endog-enous expression of connective tissue growth factor by itsantisense oligonucleotide and antisense RNA suppressesproliferation and migration of vascular endothelial cells. JBiochem 1998;124:130–40.

53 Nakanishi T, Nishida T, Shimo T, et al. EVects ofCTGF/Hcs24, a product of a hypertrophic chondro-cyte-specific gene, on the proliferation and diVerentiationof chondrocytes in culture. Endocrinology 2000;141:264–73.

54 Nishida T, Nakanishi t, Shimo T, et al. Demonstration ofreceptors specific for connective tissue growth factor on ahuman chondrocytic cell line (HCS-2/8). Biochem BiophysRes Commun 1998;247:905–9.

55 Nishida T, Nakanishi T, Asano M, et al. EVects ofCTGF/Hcs24, a hypertrophic chondrocyte-specific geneproduct, on the proliferation and diVerentiation of osteob-lastic cells in vitro. J Cell Physiol 2000;184:197–206.

56 Barbot W. DEA biologie cellulaire et moleculaire. UniversitéParis VI. 1997.

57 Perbal B. Caractérisation et expression du proto-oncogènenov humain dans les tumeurs de Wilms. Bull Cancer (Paris)1994;81:957–61.

58 Yang GP, Lau LF. Cyr61, product of a growth factor-inducible immediate early gene, is associated with theextracellular matrix and the cell surface. Cell Growth DiVer1991;2:351–7.

59 Martinerie C, HuV V, Joubert I, et al. Structural analysis ofthe human nov proto-oncogene and expression in Wilmstumor. Oncogene 1994;9:2729–32.

60 Lopez-Bermejo A, Buckway CK, Devi GR, et al. Characteri-zation of insulin-like growth factor-binding protein-relatedproteins (IGFBP-rPs) 1, 2 and 3 in human prostate epithelialcells: potential roles for IGFBP-rP1 and 2 in senescence ofthe prostatic epithelium. Endocrinology 2000;141:4072–80.

61 Li WC, Martinerie C, Zumkeller W, et al. DiVerentialexpression of novH and CTGF in human glioma cell lines.J Clin Pathol: Mol Pathol 1996;49:M91–7.

62 Babic A, Kireeva M, Kolesnikova T, et al. CYR61, a productof a growth factor-inducible immediate early gene,promotes angiogenesis and tumor growth. Proc Natl AcadSci U S A 1998;95:6355–60.

63 Babic A, Chen CC, Lau L. Fisp 12/mouse connective tissuegrowth factor mediates endothelial cell adhesion andmigration through integrin ávâ3, promotes endothelial cellsurvival, and induces angiogenesis in vivo. Mol Cell Biol1999;19:2958–66.

64 Martinerie C, Viegas-Pequignot E, Van-Cong N, et al.Chromosomal mapping and expression of the humanCYR61 gene in tumor cells from the nervous system. J ClinPathol: Mol Pathol 1997;50:310–17.

65 Gemini M, Schwalbe P, Scholl FA, Schufer BW. Isolation ofgenes diVerentially expressed in human primary myoblastsand embryonal rhabdomyosarcoma. Int J Cancer 1996;66:571–7.

66 Pilarsky C, Schmidt U, Eissrich C, et al. Expression of theextracellular matrix signaling molecule Cyr61 is downregu-lated in prostate cancer. Prostate 1998;36:85–91.

67 Tsai MS, Hornby AE, Lakins J, et al. Expression and func-tion of CYR61, an angiogenic factor, in breast cancer celllines and tumor biopsies. Cancer Res 2000;60:5603–7.

68 Frazier K, Grotendorst G. Expression of connective tissuegrowth factor mRNA in the fibrous stroma of mammarytumors. Int J Biochem Cell Biol 1997;29:153–61.

69 Igarashi A, Hayashi N, Nashiro K, et al. DiVerential expres-sion of connective tissue growth factor gene in cutaneousfibrohistiocytic and vascular tumor. J Cutan Pathol1998;25:143–8.

70 Wenger C, Ellenrieder V, Alber B, et al. Expression and dif-ferential regulation of connective tissue growth factor inpancreatic cancer cells. Oncogene 1999;18:1073–80.

71 Shakunaga T, Ozaki T, Ohara N, et al. Expression ofconnective tissue growth factor in cartilaginous tumors.Cancer 2000;89:1466–73.

72 Kumar S, Hand AT, Connor JR, et al. Identification andcloning of a connective tissue growth factor-like cDNAfrom human osteoblasts encoding a novel regulator of oste-oblast functions. J Biol Chem 1999;274:17123–31.

73 Fan WH, Karnovsky MJ. Activation of protein kinase Cinhibits the expression of connective tissue growth factor.Biochem Biophys Res Commun 2000;275:312–21.


www.molpath.com

on Decem


http://mp.bm

j.com/

Mol P


pril 2001. Dow

nloaded from

http://mp.bmj.com/

review nov (nephroblastoma overexpressed) and the ccn ... · (nephroblastoma overexpressed gene)....

Documents