Mutations in the CCN gene family member WISP3 cause progressive ...

letter

© 1999 Nature America Inc. • http://genetics.nature.com

Mutations in the CCN gene family member WISP3 cause progressive pseudorheumatoid dysplasia


Jennifer R. Hurvitz1, Wafaa M. Suwairi1,2,3, Wim Van Hul4, Hatem El-Shanti5, Andrea Superti-Furga6, Jean Roudier7, Daniel Holderbaum8, Richard M. Pauli9, J. Kenneth Herd10, Els Van Hul4, Hossien Rezai-Delui11, Eric Legius12, Martine Le Merrer13, Jamil Al-Alami14, Sultan A. Bahabri3 & Matthew L. Warman1

Members of the CCN (for CTGF, cyr61/cef10, nov) gene family encode cysteine-rich secreted proteins with roles in cell growth and differentiation1. Cell-specific and tissue-specific differences in the expression and function of different CCN family members suggest they have non-redundant roles. Using a positional-candidate approach, we found that mutations in the CCN family member WISP3 are associated with the autosomal recessive skeletal disorder progressive pseudorheumatoid dysplasia (PPD; MIM 208230). PPD is an autosomal recessive disorder that may be initially misdiagnosed as juvenile rheumatoid arthritis2–5. Its population incidence has been estimated at 1 per million in the United Kingdom4, but it is likely to be higher in the Middle East and Gulf States6. Affected individuals are asymptomatic in early childhood2,3. Signs and symptoms of disease typically develop between three and eight years of age. Clinically and radiographically, patients experience continued cartilage loss and destructive bone changes as they age2–7, in several instances necessitating joint replacement surgery by the third decade of life. Extraskeletal manifestations have not been reported in PPD. Cartilage appears to be the primary affected tissue, and in one patient, a biopsy of the iliac crest revealed abnormal nests of chondrocytes and loss of normal cell columnar organization in growth zones5. We have identified nine different WISP3 mutations in unrelated, affected individuals, indicating that the gene is essential for normal post-natal skeletal growth and cartilage homeostasis.

Typically, PPD symptoms consist of stiffness and swelling of joints, motor weakness and joint contractures2−5 (Fig. 1). The hands are affected first, then the knees, hips, spine and other large joints as the disease progresses2–7. Loss of articular cartilage continues after skeletal growth is completed. In older children and adults, radiographic examination distinguishes PPD from rheumatoid arthritis by showing loss of joint space, widened epiphyses and vertebral flattening2−7 (Fig. 1). PPD was mapped to a 3-cM interval on chromosome 6q22 by linkage studies in consanguineous families8,9. We studied 10 additional families and narrowed the candidate interval to 2 cM, between D6S1594 and D6S1706 (data not shown). The fully linked marker D6S416 is contained within this interval. D6S416 is also present in the completely sequenced PAC dJ142L7. This PAC sequence contains exons for two genes, LAMA4 and

WISP3. WISP3 is a member of the CCN family, which encode putative growth regulators10. The high conservation of cysteines in the CCN proteins suggests that these residues are essential for function1,10. Little is known about the expression pattern and

a

c

b

d

Fig. 1 Clinical and radiographic findings in a 13-year-old male PPD patient. a, Left hand. Note enlargement of the proximal interphalangeal joints (arrows). b, Hand radiograph demonstrating enlargement of the epiphyseal and metaphyseal portions of the metacarpals and phalanges. Arrows point to the same proximal interphalangeal joints as in (a). c, Knee radiograph demonstrating enlargement of the femoral and tibial epiphyses and a reduction in cartilage thickness indicated by the narrowing of the joint space (arrows). d, Lateral spine radiograph demonstrating flattening and anterior beaking (arrow) of the thoraco-lumbar spine.

1Department of Genetics and Center for Human Genetics, Case Western Reserve University School of Medicine and University Hospitals of Cleveland, Cleveland, Ohio 44106, USA. 2Department of Pediatrics, Royal Military Hospital and 3King Faisal Specialist Hospital and Research Center, Riyadh, Saudi Arabia. 4Department of Medical Genetics, University of Antwerp, Antwerp, Belgium. 5Departments of Pediatrics and Medical Laboratory Sciences and 14Departments of Biochemistry and Medical Laboratory Sciences, Jordan University of Science and Technology, Irbid, Jordan. 6Division of Molecular Pediatrics, Department of Pediatrics, University of Zurich, Zurich, Switzerland. 7Immunorheumatology, INSERM E9940, Université de la Méditerranée, Marseille, France. 8Division of Rheumatology, Department of Medicine, University Hospitals of Cleveland, Cleveland, Ohio, USA. 9Division of Genetics, Children’s Hospital, University of Wisconsin, Madison, Wisconsin, USA. 10Department of Pediatrics, East Tennessee State University, James H. Quillen College of Medicine, Johnson City, Tennessee, USA. 11Department of Radiology, Mashad University, Ghaen Hospital, Mashad, Iran. 12Centre for Human Genetics, Catholic University of Leuven, Leuven, Belgium. 13Department of Genetics and INSERM U393, Hospital Necker Enfants Malades, Paris, France.

Correspondence should be addressed to M.L.W. (e-mail: [email protected]).

94

nature genetics • volume 23 • september 1999

letter



Fig. 2 Schematic depicting WISP3 protein domains and the mutations identified in PPD patients. a, Sequencing gels demonstrating the compound heterozygous WISP3 mutations found in family 3. Sequence is derived from subclones containing wild-type or mutant patientderived PCR amplimer. The direction of translation is indicated by the vertical arrow above the wild-type codon. b, Locations of the WISP3 mutations that are likely to cause disease. Families in which they were found are indicated by circled numbers. The sites of alteration relative to the predicted unprocessed protein are indicated by arrows. Numbering for the missense and nonsense mutations reflects amino acid residues. Numbering for the frameshift mutations (italics) reflects nucleotide residues. The two homozygous frameshift mutations are underlined. The homozygous insT+2IVS1 mutation present in family 8 is not shown. Domains having homology to non-CCN protein families are indicated (IGF-BP, insulin-like growth factor binding protein; VWC, Von Willebrand type C repeat; THBS, thrombospondin type 1 repeat; CK, cysteine knot).

a

b

function of WISP3 (ref. 10), but other CCNs (refs 10−16) appear to be secreted, matrix-bound or membrane-bound proteins12,13,17 involved in cell growth and differentiation15,16,18–24. IGFBP10 (formerly Cyr61) is expressed at sites of mesenchymal cell differentiation into chondrocytes18,21. Because skeletal growth is dependent not only on the differentiation of mesenchymal cells into chondrocytes, but also on chondrocyte proliferation, differentiation and homeostasis25, we evaluated WISP3 as a candidate gene for PPD.

We sequenced all 5 coding exons of WISP3 in 13 unrelated individuals affected with PPD and found 11 sequence alterations. Two are likely to be benign polymorphisms. Nine are likely to be disease-causing as they result in frameshifts, nonsense mutations, non-synonymous changes involving cysteines, or affect a splice-donor site (Fig. 2 and Table 1). The mutations are homozygous in patients from consanguineous families. In two Saudi Arabian families that share a common disease-associ-

Table 1 • WISP3 mutation detection in families with PPD Family

Origin

Family structurea

DNA changeb

Zygosityd

Controlse

1

Italy

NC

434G→A 993G→A

cys145tyr trp331stop

Protein changec

compound heterozygous

0.00 0.00

2

Italy

NC

156C→A

cys52stop

heterozygous

0.00

3

France

NC

156C→A 232T→C

cys52stop cys78arg

compound heterozygous

0.00

4 5 9

Saudi Arabia Saudi Arabia Jordan

C C C

246delA 246delA 246delA

frameshift at residue 82 with stop 21 residues downstream

homozygous homozygous homozygous

0.00 0.00 0.00

7

France

NC

none detected

8

Jordan

C

insT+2IVS1

alters intron 1 splice donor site

homozygous

0.00

10

Jordan

C

none detected

11

USA

NC

863insAC 43delGC

12

Iran

13 6 6,7,10,11,13

homozygousf frameshift at residue 288 with stop 24 residues downstream frameshift at residue 15 with stop 15 residues downstream

0.00

heterozygous

0.00

homozygous

0.00

C

866delAG

Belgium

NC

none detected

USA

NC

807A→Gg

gln269glng

0.02

see above

168G→Tg

gln56hisg

0.25

see above

frameshift at residue 289 with stop 11 residues downstream

compound

aNC

= non-consanguineous; C = consanguineous. bNumbering begins with the methionine translation initiation codon. cNumbering begins with the methionine residue. dMutant alleles in affected patients are either homozygous, compound heterozygous or heterozygous; all putative deleterious alleles were inherited from asymptomatic carrier parents. eFrequency of allele among 100 unaffected control alleles; controls comprise 50 healthy unrelated persons of U.S. origin. fAffected family members are homozgyous for linked markers across the interval containing WISP3, but do not have detectable coding sequence mutations. gLikely benign polymorphism.


95

letter



Fig. 3 WISP3 is expressed in skeletally derived cells. The ethidium bromide stained 3% agarose gel depicts products of amplification by RT-PCR from 5 different cell types for 4 genes. PCR reactions using each primer pair (COL11A2, WISP3, CACP and UBE1) were performed separately on reverse transcribed total RNA and then pooled before electrophoresis. All primer pairs span large introns; only products derived from the spliced transcripts are present. Lanes: low molecular weight standard; negative control (RT-PCR using a water-only template); EBV-transformed lymphoblast RNA template; primary human fibroblast RNA template; primary human synoviocyte RNA template; primary human chondrocyte RNA template; bone marrow-derived mesenchymal progenitor cell RNA from day 14 of chondrogenic differentiation. COL11A2 encodes the α2 chain of type XI collagen, a quantitatively minor cartilage collagen. COL11A2 is also transcribed in EBV-transformed lymphoblasts and fibroblasts. CACP encodes the protein defective in the camptodactyly-arthropathy-coxa vara-pericarditis syndrome (manuscript submitted). UBE1 encodes ubiqitin-activating enzyme E1, a putative ‘housekeeping’ protein.

ated haplotype across the candidate interval, the affected patients are also homozygous for an identical deletion (246delA). The same deletion was found in a Jordanian patient who has a different linked haplotype. This result suggests the mutation is either very old or has arisen independently in the different kindreds. Patients from non-consanguineous families all appear to be compound heterozygotes. We were able detect both mutant alleles in 8 of 13 families. In one family, we identified only one of two mutant alleles. The occurrence of this mutation in an asymptomatic heterozygous parent excludes its being a de novo mutation with a dominant effect; the second mutant allele is presumably in non-coding sequence. We have been unable to RT-PCR amplify WISP3 from patient or control EBV-transformed lymphoblasts and cannot perform mutation detection with any currently available mRNA source. We were unable to identify disease-causing mutations in four families. Three of them are too small to determine whether the phenotype is linked to WISP3. Locus heterogeneity or phenocopy may therefore account for our failure to find mutations in some families; however, the fourth family in which we have not yet found a WISP3 mutation is consanguineous, and affected individuals are homozygous for genetic markers across the interval that contains WISP3. The identification of nine independent WISP3 mutations in unrelated PPD patients suggests a causative role for this gene. It seems likely that the mutations we have identified abrogate the protein’s normal function in the skeleton, rather than create a new or interfering function, because all obligate carriers are asymptomatic. This also suggests that functional haploinsufficiency for the WISP3 product is tolerated. CCN family members have a modular architecture with putative functional domains roughly corresponding to each exon1. Typically, the peptide signal sequence is encoded by exon 1. Exon 2 encodes an amino-terminal domain that has homology to the insulin-like growth factor binding proteins (IGF-BP). Exon 3 encodes a domain with homology to Von Willebrand factor type C (VWC) repeats and may participate in peptide oligomerization. WISP3, unlike all other CCN proteins identified to date, lacks 4 of 10 conserved cysteine residues in this domain10. Exon 4 encodes a thrombospondin type I domain and may be involved in the binding of CCNs to sulfated glycosaminoglycans either at cell surfaces or in extracellular matrix. Exon 5 encodes a ‘cysteine knot’ domain, which has been identified in several other signalling peptides (such as transforming growth factor β, platelet derived growth factor and nerve growth factor), and may participate in dimerization and receptor binding. 96

The identification of patients who are homozygous for frameshift mutations in WISP3 may be used to predict the likely consequence of each mutation’s effect at the protein level. It has been suggested that the different functional domains of CCNs can act independently of each other1. For example, WISP2 completely lacks the ‘cysteine knot’-containing domain, yet still regulated cell growth10,15, whereas fragments of connective tissue growth factor (CTGF) that retain their ‘cysteine knot’ motif remain mitogenic26. Furthermore, a truncated form of NOV was capable of inducing cell transformation in vitro14. In contrast, WISP3 frameshift and nonsense mutations in exons 1, 2 and 5 seem to cause identical clinical features, indicating that synthesis of in-frame, full-length WISP3 transcript is essential for either mRNA or polypeptide stability or for WISP3 protein function. Disease-associated missense mutations at conserved cysteines in both the IGF-BP domain and the VWC domain imply that proper folding of all protein domains is also essential. WISP3 has two closely related homologues, WISP1 and WISP2, which were identified through their increased expression in Wnt1 transformed cells10. The Wnt family of signalling proteins also regulate cell fate, motility, morphology and proliferation27 and are important mediators of skeletal development25,28. WISP1 and WISP2 also demonstrate differential expression in primary human colon cancers. WISP3 was identified by database searching10. Although WISP3 is differentially expressed in primary human colon tumours relative to its expression in the patient’s normal mucosa10, the physiological relevance of this observation is unknown. The lack of extraskeletal manifestations in patients with PPD suggests that WISP3’s primary non-redundant role is in the regulation of post-natal skeletal growth and long-term skeletal homeostasis. Although we can easily detect WISP3 genomic DNA by Southern blot, we have been unable to detect transcripts with commercially available mouse northern blots, northern blots containing RNA extracted from mouse long bones, vertebrae and rib cartilage, or multi-tissue adult human northern blots, even after long exposures. In addition, few WISP3 sequences are contained in the human and mouse EST databases. This suggests WISP3 is likely to have a low level of expression. Expression of WISP2 (formerly r-Cop1), another CCN family member, has also been difficult to detect15. Using RT-PCR, WISP3 cDNA has been amplified from tissues including kidney, testis, placenta, ovary, prostate and small intestine10. Consistent with its role in skeletal growth and homeostasis, we observed WISP3 expression by RT-PCR in skeletally derived cells including human synoviocytes, articular cartilage chondrocytes and bonemarrow-derived mesenchymal progenitor cells which have been induced to undergo chondrogenesis in vitro29 (Fig. 3). nature genetics • volume 23 • september 1999


Crucial to the delineation of the physiological role of WISP3 is our understanding of the disorder PPD. In contrast to most genetic skeletal dysplasias, pre-natal skeletal growth and morphogenesis appears undisturbed in PPD, and affected individuals are asymptomatic during the first years of life. In two affected children, skeletal radiographs taken at three years of age did not detect an underlying skeletal dysplasia2,3. Pain and swelling of joints occurs during childhood. As symptoms progress, radiographic changes in cartilage and bone become apparent. The early clinical picture of PPD resembles rheumatoid arthritis; this is commonly the initial diagnosis, although patients lack most signs of inflammation. The course of PPD is that of severe and painful joint degeneration. Thus, absence of WISP3 apparently interferes with normal post-natal growth and cartilage homeostasis, leading to precocious joint degeneration.


Methods Clinical material. After obtaining informed consent, blood or cheek swab DNA was extracted from 13 families affected by PPD. Of the 13 families, 6 were consanguineous and 7 were non-consanguineous. Patients were diagnosed as being affected with PPD based on published clinical and radiographic criteria2−7. Clinical features common to all patients included onset of symptoms in early childhood, enlargement of the proximal interphalangeal joints and progressive stiffness and contractures of other joints. Radiographic features common to all patients were platyspondyly, widened epiphyses and metaphyses, joint space narrowing and cartilage loss. Clinical descriptions for three of the affected kindreds have been published: families 9 (ref. 2), 12 (ref. 3) and 13 (ref. 30). All obligate heterozygous carriers deny symptoms of early onset arthritis, osteoporosis or joint contractures. Cells used in RT-PCR assays of WISP3 expression were obtained, after obtaining informed consent, from non-PPD patients who had undergone knee replacement surgery for osteoarthritis. Identification of WISP3 as a candidate gene. We performed BLAST (ref. 31) searches using the sequences of all simple sequence repeat-containing markers in the PPD candidate region to identify sequenced large insert clones in GenBank. Only D6S416 identified such a clone, PAC dJ142L7. The PAC sequence was then edited with RepeatMasker (http://ftp.genome.washington.edu/RM/RepeatMasker.html) and used to BLAST the non-redundant and EST databases. Two potential candidates were identified. One was LAMA4, the other was an EST (IMAGE clone 1076664) with homology to the CCN gene family. This EST contains part of WISP3. We deduced the intron-exon boundaries of WISP3 by comparing the full-length WISP3 cDNA sequence10 to the genomic sequence from the PAC. Exon 1 is not contained within the PAC sequence. Identification of genomic sequence containing exon 1. We probed the RPCI-11 Human BAC Library (Research Genetics) with WISP3 exon 2 amplimer and identified the positive clone 10D17. BAC DNA from this clone was isolated using the Nucleobond kit (Clontech). Exons 2−5 of WISP3 were amplified using this BAC DNA as template, confirming that it contains WISP3. We used a primer designed from the 5′ end sequence of PAC 142L7 (in intron 1 of WISP3) and the 1F3R forward primer (Table 2, published electronically only at http://genetics.nature.com/supplementary_info/) on BAC 10D17 DNA to amplify a 2.8-kb fragment which contains exon 1 and the 5′ portion of intron 1. Sequencing the ends of this amplimer permitted us to design a reverse primer nearer to exon 1, located 3′ of the intron 1 splice donor site (Table 2). Direct sequencing of BAC 10D17 using this primer allowed the design of a forward primer, located 5′ of WISP3’s putative translation initiation codon (Table 2).


letter

Amplification and sequencing of WISP3 from families with PPD and controls. We designed a primer pair to amplify exon 1 and chose intronic primers to amplify exons 2−5 using the Primer3 program (http://wwwgenome.wi.mit.edu/genome_software/other/primer3.html; primer sequences available at http://genetics.nature.com/supplementary_info). PCR amplification was carried out in 50-µl reactions containing 50 ng genomic DNA, 0.2 µM each primer and 200 µM each dNTP. Cycling conditions included 35 cycles of 95 °C for 30 s, 56 °C for 45 s and 72 °C for 50 s, with an initial 4-min denaturation step and a final 10-min extension. To amplify exon 4, we needed to denature the reaction mix at 95 °C before adding Taq polymerase. We purified PCR products using the Wizard PCR prep kit (Promega) and sequenced the purified amplimers with the fmol DNA cycle sequencing system (Promega) using 33P-labelled internal primers (primer sequences available on request). We demonstrated that patients with PPD who had two different mutations were compound heterozygotes either by subcloning and sequencing individual amplimers or by demonstrating that each mutation was inherited from a different parent. None of the 9 likely disease causing mutations were observed in 100 control alleles. Northern- and Southern-blot analysis. We probed a human multiple tissue northern blot (Clontech) and a human genomic DNA Southern blot with a 276-bp fragment within exon 5 of WISP3. A mouse multiple embryonic stage northern blot and a northern blot containing 5 µg of total RNA each from mouse long bone, vertebrae, rib cartilage, liver and kidney were probed with the homologous 276-bp fragment of mouse Wisp3. The amplimers were labelled with 32P dCTP using the High Prime labelling kit (Boehringer) and hybridized at 68 °C (northern) and 60 °C (Southern) in ExpressHyb hybridization solution (Clontech). The blots were washed according to the manufacturer’s recommendation, and exposed to a phosphor screen. Images of the blots were viewed with ImageQuant software. RT-PCR. To demonstrate expression and splicing of WISP3 mRNA, we used primer pair 1F3R (Table 2) on cell-derived cDNA. Expression of other mRNAs including COL11A2 (ref. 32; primer pair 1/6), UBE1 (ref. 33; primer pair1/7) and CACP (primer pair Q, manuscript submitted) were used as controls. Cycling conditions for WISP3, UBE1 and CACP included 35 cycles of 95 °C for 30 s, 56 °C for 45 s and 72 °C for 50 s, with an initial 4min denaturation step and a final 10-min extension; COL11A2 was amplified as described32. RT-PCR was performed on 2 µl of cDNA made from total RNA recovered from EBV-transformed lymphoblasts, primary human cultures of skin fibroblasts, knee joint synoviocytes and knee joint chondrocytes, and from iliac crest bone-marrow-derived mesenchymal progenitor cells (at day 14) after in vitro chondrogenesis had been induced29. RNA was extracted by standard methods and cDNA was prepared from total RNA using the superscript preamplification system (GibcoBRL). PCR reactions (25 µl) contained 2 µl template cDNA, 0.2 µM each primer, and 200 µM each dNTP. GenBank accession numbers. PAC dJ142L7, Z99289; IMAGE clone 1076664, AA592984; WISP3, AF100781. Acknowledgements

We thank the patients and their families for participating in this study, and J. Marcelino, Y. Gong, S. Gregory, P. Modaff, T. Haqqi, B. Johnstone, F.M. Pope and A. Richards for sharing clinical and scientific expertise. This work was supported by NIH grant AR43827 and by a biomedical sciences grant from the March of Dimes Birth Defects Foundation (M.L.W.), grant 45-401.95 from the Swiss National Foundation (A.S.-F.) and a concertedaction grant from the University of Antwerp (W.V.H.).

Received 24 May; accepted 22 July 1999.

97

letter 1. 2. 3.

4.

5.

6.

7. 8.

9. 10.

11.


12.

13.

14.

15.

16.

98


Bork, P. The modular architecture of a new family of growth regulators related to connective tissue growth factor. FEBS Lett. 327, 125–130 (1993). El-Shanti, H.E., Omari, H.Z. & Qubain, H.I. Progressive pseudorheumatoid dysplasia: report of a family and review. J. Med. Genet. 34, 559–563 (1997). Rezai-Delui, H., Mamoori, G., Sadri-Mahvelatie, E. & Noori, N.M. Progressive pseudorheumatoid chondrodysplasia: a report of nine cases in three families. Skeletal Radiol. 23, 411–419 (1994). Wynne-Davies, R., Hall, C. & Ansell, B.M. Spondylo-epiphysial dysplasia tarda with progressive arthropathy. A “new” disorder of autosomal recessive inheritance. J. Bone Joint Surg. Br. 64, 442–445 (1982). Spranger, J., Albert, C., Schilling, F., Bartsocas, C. & Stöss, H. Progressive pseudorheumatoid arthritis of childhood (PPAC). A hereditary disorder simulating rheumatoid arthritis. Eur. J. Pediatr. 140, 34–40 (1983). Teebi, A.S. & Al-Awadi, S.A. Spondyloepiphyseal dysplasia tarda with progressive arthropathy: a rare disorder frequently diagnosed among Arabs. J. Med. Genet. 23, 189–191 (1986). Kozlowski, K., Kennedy, J. & Lewis, I.C. Radiographic features of progressive pseudorheumatoid arthritis. Australas. Radiol. 30, 244–250 (1986). El-Shanti, H. et al. Assignment of gene responsible for progressive pseudorheumatoid dysplasia to chromosome 6 and examination of COL10A1 as candidate gene. Eur. J. Hum. Genet. 6, 251–256 (1998). Fischer, J. et al. Genetic linkage of progressive pseudorheumatoid dysplasia to a 3cM interval of chromosome 6q22. Hum. Genet. 103, 60–64 (1998). Pennica, D. et al. WISP genes are members of the connective tissue growth factor family that are up-regulated in Wnt-1-transformed cells and aberrantly expressed in human colon tumors. Proc. Natl Acad. Sci. USA 95, 14717–14722 (1998). O’Brien, T.P., Yang, G.P., Sanders, L. & Lau, L.F. Expression of cyr61, a growth factor-inducible immediate-early gene. Mol. Cell. Biol. 10, 3569–3577 (1990). Ryseck, R.-P., Macdonald-Bravo, H., Mattei, M.-G. & Bravo, R. Structure, mapping, and expression of fisp-12, a growth factor-inducible gene encoding a secreted cysteine-rich protein. Cell Growth Differ. 2, 225–233 (1991). Bradham, D.M., Igarashi A., Potter, R.L. & Grotendorst, G.R. Connective tissue growth factor: a cysteine-rich mitogen secreted by human vasular endothelial cells is related to the SRC-induced immediate early gene product CEF-10. J. Cell Biol. 114, 1285–1294 (1991). Joliot, V. et al. Proviral rearrangements and overexpression of a new cellular gene (nov) in myeloblastosis-associated virus type 1-induced nephroblastomas. Mol. Cell. Biol. 12, 10–21 (1992). Zhang, R. et al. Identification of rCop-1, a new member of the CCN protein family, as a negative regulator for cell transformation. Mol. Cell. Biol. 18, 6131–6141 (1998). Hashimoto, Y. et al. Expression of the Elm1 gene, a novel gene of the CCN (connective tissue growth factor, Cyr61/Cef10, and neuroblastoma overexpressed gene) family, suppresses in vivo tumor growth and metastasis of K-1735 murine melanoma cells. J. Exp. Med. 187, 289–296 (1998).

17. Yang, G.P. & Lau, L.F. Cyr61, product of a growth factor-inducible immediate early gene, is associated with the extracellular matrix and the cell surface. Cell Growth Differ. 2, 351–357 (1991). 18. O’Brien, T.P. & Lau, L.F. Expression of the growth factor-inducible immediate early gene cyr61 correlates with chondrogenesis during mouse embryonic development. Cell Growth Differ. 3, 645–654 (1992). 19. Kireeva, M.L., Mo, F., Yang, G.P. & Lau, L.F. Cyr61, a product of a growth factorinducible immediate-early gene, promotes cell proliferation, migration, and adhesion. Mol. Cell. Biol. 16, 1326–1334 (1996). 20. Martinerie, C., Chevalier, G., Rauscher, F.J. III & Perbal, B. Regulation of nov by WT1: a potential role for nov in nephrogenesis. Oncogene 12, 1479–1492 (1996). 21. Wong, M., Kireeva, M.L., Kolesnikova, T.V. & Lau, L.F. Cyr61, a product of a growth factor-inducible immediate-early gene, regulates chondogenesis in mouse limb bud mesenchymal cells. Dev. Biol. 192, 492–508 (1997). 22. Grotendorst, G.R. Connective tissue growth factor: a mediator of TGF-β action on fibroblasts. Cytokine Growth Factor Rev. 8, 171–179 (1997). 23. Kothapalli, D., Frazier, K.S., Welply, A., Segarini, P.R. & Grotendorst, G.R. Transforming growth factor B induces anchorage-independent growth of NRK fibroblasts via a connective tissue growth factor-dependent signaling pathway. Cell Growth Differ. 8, 61–68 (1997). 24. Schütze, N. et al. The human analog of murine cystein rich protein 16 (sic) is a 1∝,25-dihydroxyvitamin D3 responsive immediate early gene in human fetal osteoblasts: regulation by cytokines, growth factors, and serum. Endocrinology 139, 1761–1770 (1998). 25. Erlebacher, A., Filvaroff, E.H., Gitelman, S.E. & Derynck, R. Toward a molecular understanding of skeletal development. Cell 80, 371–378 (1995). 26. Brigstock, D.R. et al. Purification and characterization of novel heparin-binding growth factors in uterine secretory fluids. J. Biol. Chem. 272, 20275–20282 (1997). 27. Dale, T.C. Signal transduction by the Wnt family of ligands. Biochem. J. 329, 209–223 (1998). 28. Parr, B.A., Avery, E.J., Cygan, J.A. & McMahon, A.P. The classical mouse mutant postaxial hemimelia results from a mutation in the Wnt 7a gene. Dev. Biol. 202, 228−234 (1998). 29. Yoo, J.U. et al. The chondrogenic potential of human bone-marrow-derived mesenchymal progenitor cells. J. Bone Joint Surg. Am. 80, 1745–1757 (1998). 30. Legius, E., Mulier, M., Van Damme, B. & Fryns, J.P. Progressive pseudorheumatoid arthritis of childhood (PPAC) and normal adult height. Clin. Genet. 44, 152–155 (1993). 31. Altschul, S.F., Gish, W., Miller, W., Myers, E.W. & Lipman, D.J. Basic local alignment search tool. J. Mol. Biol. 215, 403–410 (1990). 32. Vikkula, M. et al. Autosomal dominant and recessive osteochondrodysplasias associated with the COL11A2 locus. Cell 80, 431–437 (1995). 33. Carrel, L. et al. X inactiviation analysis and DNA methylation studies of the ubiquitin activating enzyme E1 and PCTAIRE-1 genes in human and mouse. Hum. Mol. Genet. 5, 391–401 (1996).
