ccn2 is necessary for the function of mouse embryonic fibroblasts

E X P E R I M E N T A L C E L L R E S E A R C H 3 1 3 ( 2 0 0 7 ) 9 5 2 – 9 6 4

ava i l ab l e a t www.sc i enced i r ec t . com

www.e l sev i e r. com/ loca te /yexc r

Research Article

CCN2 is necessary for the function of mouseembryonic fibroblasts

Laura Kennedya, Shangxi Liua, Xu Shi-wenb, Yunliang Chenb,c,Mark Eastwoodc, David E. Carterd, Karen M. Lyonse, Carol M. Blackb,David J. Abrahamb, Andrew Leaska,⁎aDivision of Oral Biology and Department of Physiology and Pharmacology, CIHR Group in Skeletal Development and Remodeling,Schulich School of Medicine and Dentistry, University of Western Ontario, Dental Sciences Building, London, ON, Canada N6A 5C1bCentre for Rheumatology and Connective Tissue Disease, University College London (Royal Free Campus), Rowland Hill Street,London NW3 2PF, UKcSchool of Biosciences, University of Westminster, 115 New Cavendish Street, London, W1W 6UW, UKdLondon Regional Genomics Centre, Robarts Building, London, ON, Canada N6A 5K8eDepartments of Biological Chemistry and Orthopedic Surgery, David Geffen Medical School at UCLA, University of California,Los Angeles, CA 90095, USA

A R T I C L E I N F O R M A T I O N

⁎ Corresponding author.E-mail address: [email protected]

0014-4827/$ – see front matter © 2006 Elsevidoi:10.1016/j.yexcr.2006.12.006

A B S T R A C T

Article Chronology:Received 18 September 2006Revised version received15 December 2006Accepted 18 December 2006Available online 23 December 2006

CCN2 is expressed by mesenchymal cells undergoing active tissue remodeling, and ischaracteristically overexpressed in connective tissue pathologies such as fibrosis andcancer. However, the physiological roles and mechanism of action of CCN2 are largelyunknown. Here, we probe the contribution of CCN2 to the biology of mouse embryonicfibroblasts (MEFs) using genome-widemRNA expression profiling, proteomic and functionalbioassay analyses. We show that ccn2−/− mouse embryonic fibroblasts (MEFs) havesignificantly reduced the expression of pro-adhesive, pro-inflammatory and pro-angiogenicgenes such as interleukin-6 (IL-6), ceruloplasmin, thrombospondin-1, lipocalin-2 andsyndecan 4. Anti-syndecan 4 antibody reduced ERK phosphorylation in ccn2+/+ MEFs. Inccn2+/+ MEFs, the MEK inhibitor U0126 and dominant negative ras reduced expression ofIL-6 and lipocalin-2. Overexpressing syndecan 4 in ccn2−/−MEFs restored IL-6 and lipocalin-2 mRNA expression. Syndecan 4 has been shown to mediate cell migration. We found thatccn2+/+ MEFs migrated significantly faster than ccn2−/− MEFs; anti-syndecan 4 antibodyand U0126 reduced the migration of ccn2+/+ MEFs to that of ccn2−/− MEFs. These resultscollectively support the notion that syndecan 4 acts downstream of CCN2 inMEFs, and thatreduced syndecan 4 expression contributes to at least part of the ccn2−/− phenotype.Further, these results suggest that CCN2 is required for MEFs to contribute to aspects oftissue remodeling. Consistent with this notion, whereas ccn2+/+ MEFs displayed actinstress fibers and focal adhesions at the cell periphery consistent with a migratoryphenotype, ccn2−/− MEFs displayed reduced focal adhesions and actin stress fibers, and areduced ability to transduce forces across a collagen gel matrix. Collectively, these results

Keywords:CCN2Embryonic fibroblastsTissue remodelingConnective tissue growth factor

wo.ca (A. Leask).

er Inc. All rights reserved.

mailto:[email protected]

http://dx.doi.org/10.1016/j.yexcr.2006.12.006

953E X P E R I M E N T A L C E L L R E S E A R C H 3 1 3 ( 2 0 0 7 ) 9 5 2 – 9 6 4

suggest that CCN2 supplies essential, non-redundant functions required for fibroblasts toproperly participate in features of embryogenesis, and further suggest that CCN2may playessential roles in adult wound healing, tissue repair and fibrogenesis.

© 2006 Elsevier Inc. All rights reserved.

Introduction

Tissue morphogenesis, remodeling and repair involve acomplex, coordinated program of chemotaxis, cell prolifera-tion, angiogenesis, and extracellular matrix (ECM) production[1]. CCN2 (connective tissue growth factor/CTGF), a member ofthe CCN family of proteins [2–5], is expressed inmesenchymalcells during development and wound healing [6–8] and isoverexpressed in fibrosis and cancer [9–11]. CCN2 is likely toplay a key role in these processes; however, the physiologicalrole, and mechanism of action, of CCN2, in these events areonly beginning to be uncovered.

In vitro, recombinant CCN2 acts as a potent adhesivemolecule, through integrins and heparan sulfate-containingproteoglycan (HSPG), which together act as bona fide CCN2receptors [2,12–15]. In addition to the direct effects on celladhesion, CCN2 executes other functions including cell pro-liferation, myofibroblast formation andmatrix deposition, in afashion which largely depends on the presence of co-factors,whose identity varies widely depending on the individualsituation examined [16–21]. Undoubtedly the entire range ofligands interacting with CCN family members has not beenelucidated. In addition, other CCN family members, CCN1(cyr61) and CCN3 (nov), possess similar activity to CCN2 [16,22–24], suggesting the possibility that CCN2 proteinsmay be func-tionally redundant in vivo. To gain greater molecular insightsinto the precisemechanisms of CCN action, genetic analysis ofthe roles of CCN family members is therefore essential.

To begin to investigate the physiological function of CCN2,ccn2−/− mice were recently generated [25]. These mice diesoon after birth due to ribcage malformations caused bydefects in chondrogenesis, osteogenesis and growth plateangiogenesis resulting in an inability to breathe, emphasizingthe unique contribution of CCN2 in vivo [25]. The fundamentalbasis for this phenotype is unclear, but defects in mesench-ymal cell function are likely to play a significant role. To beginto address the contribution of CCN2 to mesenchymal cellbiology in development, we have investigated the contributionof CCN2 on the function of mouse embryonic fibroblasts(MEFs), which express CCN2 constitutively, and in response toTGFβ through Smad3 and Ets-1 elements in the CCN2promoter [26–28]. Compared to ccn2+/+ MEFs, ccn2−/− MEFsshow reduced activation of adhesive signaling pathways inresponse to both fibronectin and TGFβ [15,21]. These resultssuggest that, in mesenchymal cells undergoing active matrixremodeling such as MEFs, CCN2 is required for the ability ofgrowth factors and matrix to maximally induce adhesivesignaling pathways and emphasize that CCN2 plays anessential, non-redundant function. However, the physiologi-cal contribution of endogenous CCN2 to the basal activity ofMEFs has not yet been examined.

Although CCN2 directly signals through integrins andproteoglycans [2,12–15], CCN2 can indirectly cause modifica-tions in gene expression through its ability to interact with

and modify other growth factors; for example, the ability ofCCN2 to induce type I collagen depends on EGF, IGF-2 orinsulin [17,19,20]. (Indeed, it is now appreciated that many, ifnot most, of the effects of CCN2 can be interpreted as beingindirect, occurring through its ability to modify the effects ofother factors, whose identity varies widely depending on theindividual context [4].) Nonetheless, the genes whose expres-sion in fibroblasts ultimately depends on CCN2, and directlycontributes to the phenotype of ccn2−/− MEFs, is wholly un-known. This knowledge is essential to gain valuable insightsinto the mechanism of action of CCN2, and into the basis ofthe ccn2−/− phenotype. In particular, genes whose expressionin MEFs requires CCN2, and are therefore potential down-streammediators of CCN2 action, are unknown. Identificationof such genes is essential in understanding how CCN2, bothdirectly and indirectly, participates in fibroblast biology.

In this report, to uncover the role of CCN2,we subject ccn2+/+and ccn2−/− MEFs to Affymetrix gene profiling, proteomic,Western blot and real-time polymerase chain reaction (RT-PCR) analyses and functional bioassays. Our results reveal newinsights into the physiological function, and mechanism ofaction, of CCN2 to mesenchymal cell biology.

Materials and methods

Cell culture

Ccn2+/+ and ccn2−/− MEFs (E14.5) were isolated as previouslydescribed [15, 25], and cultured in high-glucose DMEM, 10%fetal calf serum, 2 mM L-glutamine, antibiotics (100 U/mlpenicillin and 100 μg/ml streptomycin) and 1 mM sodiumpyruvate (Invitrogen, Burlington, ON) at 37 °C, 5% CO2.

Western blot analysis

Cells were cultured until 90–95% confluence, cultured over-night in the DMEM, 0.5% FBS, harvested in PBS, resuspended inRIPA buffer and a cocktail of protease and phosphataseinhibitors (Roche, Laval, QC), sonicated (Fisher, Ottawa, ON),centrifuged, and proteins present in the supernatant werequantified (BCA Kit, Pierce, Rockford, IL) and placed in proteinsample buffer (BioRad; Mississauga, ON). In some cases,neutralizing anti-syndecan 4 antibody (N-19, which recog-nizes the ectodomain of syndecan 4) or IgG control (SantaCruz, Santa Cruz, CA [29, 30]) were added overnight prior toharvesting of protein extracts. Equal amounts of protein(40 μg) were subjected to SDS-PAGE, and blotted ontonitrocellulose membranes (Invitrogen; Burlington, ON) whichwere blocked in 5% BSA in Tris-buffered saline, 0.1% Tween 20(Sigma, St. Louis, MO), overnight at 4 °C. Immunoblotting wasperformed using anti-actin, anti-α-smooth muscle actin(Sigma), anti-vinculin, anti-phospho-ERK or anti-total ERK(Cell Signaling Technology, Beverly, MA) or anti-syndecan 4

954 E X P E R I M E N T A L C E L L R E S E A R C H 3 1 3 ( 2 0 0 7 ) 9 5 2 – 9 6 4

antibodies (Zymed, South San Francisco, CA), as described bythe manufacturer. Blots were then developed by sequentialincubation with appropriate biotinylated secondary antibo-dies and ABC regent (Vector Laboratories, Burlingame, CA).Proteins were detected using a luminescence kit (Amersham,UK) and X-ray film, and quantified using Gel Base/Gel-Blot Pro(Synoptics, Cambridge, UK).

Fibroblast populated collagen lattices (FPCL)

Measurement of contractile force generated within a three-dimensional, tethered FPCL was performed as describedpreviously [31–33]. Using 1×106 cells/ml of collagen gel (FirstLink, UK), we measured the force generated across thecollagen lattice with a culture force monitor which measuresforces exerted by cells within a collagen lattice [34], over 24 has fibroblasts attach, spread, migrate and differentiate intomyofibroblasts. In brief, a rectangular fibroblast seededcollagen gel was cast and floated in medium in 2% FCS inthe presence or absence of TGFβ1 (4 ng/ml), while tethered totwo flotation bars on either side of the long edges, in turnattached to a ground point at one end and a force transducer atthe other. Cell-generated tensional forces in the collagen gelare detected by the force transducer and logged into a personalcomputer. Graphical readings are produced every 15 s provid-ing a continuous output of force (Dynes: 1×10−5N) generated[34]. The cells used in these experiments were passagematched; experiments were run in parallel and three inde-pendent times. A representative trace is shown.

Quantification of cytokines

We used a Cytometric Bead Array™ (CBA) to simultaneouslyquantitate 27 different human cytokines in the culturesupernatant, as per manufacturer's instructions (BD Bios-ciences, San Diego, CA). For each sample, 50 μl of eachantibody-microparticle reagent and antibody-PE detectorwas added to 50 μl of undiluted culture supernatant. Themixture was incubated in darkness at room temperature for3 h, and then washed before data acquisition with a FACSCalibur flow cytometer. 9000 Antibody-microparticles wereacquired for analysis using CBA Array software (FCAP Arrayv1.0). Cytokines are quantified from standard curves rangingfrom 0 to 5000 pg/ml. Alternatively, a specific IL-6 ELISA(R&D Systems; Minneapolis, MN) was used, per the manu-facturer's instructions.

Proteomic analysis

Cultured cells at subconfluent healthy state were collectedin ice-cold PBS containing 1 mM PMSF, 2 mM sodiumorthovanadate, 2 mM NaF (Sigma), and complete proteaseinhibitor cocktail (Boehringer Mannheim, Germany) andthen were lysed in modified radioimmunoprecipitation(RIPA) buffer containing the above protease and phospha-tease inhibitors. Quantitation of total protein concentrationsin the cell lysates was determined by Bio-Rad Bradfordassay. Profiling of protein expression between ccn2+/+ andccn2−/− MEFs (150 μg/sample) was performed by Kinexususing the KPKS1.2A protein kinase screen (Vancouver, BC;

see http://www.kinexus.ca/KPKS.htm for details, includingthe entire list of kinases examined). Samples were quanti-fied, using densitometry, by Kinexus. Briefly, the KPKS1.2Ascreen detects the target proteins in two steps. First,molecules were separated by gel electrophoresis based ontheir molecular weights, and then were detected by theirimmunoreactivity with highly validated antibodies. Theresulting immune complexes were subjected to a multi-plexing apparatus, and the quantitation of the bands isvisualized using ECL followed by a highly sensitive imagingsystem with a 16-bit camera and a quantitation software toanalyze the chemiluminescent samples. Each sample'simmunoblot was scanned at its maximum time to ensurethat the signal from the strongest immunoreactive proteinon the immunoblot remains below saturation. It thusprovides accurate quantitation over a 2000-fold range oflinearity.

RNA quality assessment, probe preparation and Gene Chiphybridization and analysis

Microarrays and analysis were performed essentially aspreviously described [21,35,36]. All Gene Chips were pro-cessed at the London Regional Genomics Centre (RobartsResearch Institute, London, ON; http://www.lrgc.ca). RNA washarvested (Trizol, Invitrogen), quantified and quality wasassessed using the Agilent 2100 Bioanalyzer (Agilent, PaloAlto, CA) and the RNA 6000 Nano kit (Caliper Life Sciences,Mountain View, CA). Quality data were then analyzed usingthe Degradometer (www.dnaarrays.org). Biotinylated compli-mentary RNA (cRNA) was prepared from 10 μg of total RNA asper the Affymetrix GeneChip Technical Analysis Manual(Affymetrix, Santa Clara, CA). Double-stranded cDNA wassynthesized using SuperScript II (Invitrogen) and oligo(dT) 24primers. Biotin-labeled cRNA was prepared by cDNA in vitrotranscription using the Bizarre High-Yield RNA TranscriptLabeling kit (Enzo Brioche, New York, NY) incorporatingbiotinylated UTP and CTP. Fifteen micrograms of labeledcRNA was hybridized to Mouse Genome 430 2.0 Gene Chipsfor 16 h at 45 °C as described in the Affymetrix TechnicalAnalysis Manual (Affymetrix, Santa Clara, CA). Gene Chipswere stained with streptavidin–phycoerythrin, followed by anantibody solution and a second streptavidin–phycoerythrinsolution, with all liquid handling performed by a GeneChipFluidics Station 450 (Affymetrix). Gene Chips were scannedwith the Affymetrix GeneChip Scanner 3000 (Affymetrix).Signal intensities for genes were generated using GCOS1.2(Affymetrix) using default values for the Statistical Expres-sion algorithm parameters and a Target Signal of 150 for allprobe sets and a Normalization Value of 1. Normalization wasperformed in GeneSpring 7.2 (Agilent Technologies Inc.). TheRMA preprocessor was used to import data from the .cel files.Data were first transformed (measurements less than 0.01 setto 0.01) and then normalized per chip to the 50th percentile,and per gene to wild-type control samples. Experiments wereperformed twice, and fold changes were identified using theGeneSpring filter. Data presented in Table 1 are an average ofthese independent studies. The fold change between treatedand untreated samples had to be at least twofold to identify atranscript as being altered. These criteria had to be met in

http://www.kinexus.ca/KPKS.htm

http://www.lrgc.ca

http://www.dnaarrays.org

Table 1 – Significantly (p<0.005) over-representedfunctional categories of genes, whose expression isreduced >2-fold in ccn2−/− MEFs

Affymetrix ID Gene name

Cell communicationCell adhesion1416236_a_at Epithelial V-like antigen 11419015_at WNT1 inducible signaling pathway

protein 21419309_at Podoplanin1419331_at Cadherin 171421633_a_at Hyaluronan and proteoglycan link

protein 11422450_at Catenin, delta 11422640_at Protocadherin beta 91423551_at Cadherin 131426560_a_at Nephronectin1427010_s_at Laminin, alpha 51440173_x_at Selectin, platelet1442340_x_at Cysteine-rich protein 611449254_at Secreted phosphoprotein 11449328_at Lymphocyte antigen 751449827_at Aggrecan 11450377_at Similar to thrombospondin 11456397_at Cadherin 41460302_at Thrombospondin 11448793_a_at Syndecan 4

Signal transduction1416783_at Tachykinin 11417292_at Interferon gamma inducible protein 471417625_s_at Chemokine orphan receptor 11418126_at Chemokine (C–C motif) ligand 51419080_at Glial cell line-derived neurotrophic factor1419437_at Single-minded homolog 21419728_at Chemokine (C–X–C motif) ligand 51420380_at Chemokine (C–C motif) ligand 21421404_at Chemokine (C–X–C motif) ligand 151421917_at Platelet-derived growth factor receptor,

alpha1425514_at Phosphatidylinositol 3-kinase (p85 alpha)1425663_at Interleukin 1 receptor antagonist1426063_a_at GTP-binding protein1427019_at Protein tyrosine phosphatase, receptor

type Z, polypeptide 11436173_at Deleted in liver cancer 11437270_a_at Cardiotrophin-like cytokine factor 11438946_at Platelet-derived growth factor receptor,

alpha1448831_at Angiopoietin 2

Cell–cell signaling1416783_at Tachykinin 11419080_at Glial cell line-derived neurotrophic factor1423271_at Gap junction membrane channel protein

beta 21425663_at Interleukin 1 receptor antagonist

Immune response1416783_at Tachykinin 11417268_at CD14 antigen1417483_at Nuclear factor of kappa inhibitor, zeta1418037_at Complement component-4-binding

protein1418126_at Chemokine (C–C motif) ligand 51418240_at Guanylate nucleotide-binding protein 21418293_at Interferon-induced protein with

tetratricopeptide repeats 21418392_a_at Guanylate nucleotide-binding protein 4

Table 1 (continued)

Affymetrix ID Gene name

Immune response1418480_at Chemokine (C–X–C motif) ligand 71418536_at Similar to H-2 class I histocompatibility

antigen1419209_at Chemokine (C–X–C motif) ligand 11420380_at Chemokine (C–C motif) ligand 21421404_at Chemokine (C–X–C motif) ligand 151421812_at TAP-binding protein1422962_a_at Proteosome subunit, beta type 81423954_at Complement component 31424948_x_at Histocompatibility 2, D region locus 11425336_x_at Histocompatibility 2, K1, K region1425514_at Phosphatidylinositol 3-kinase

(p85 alpha)1425663_at Interleukin 1 receptor antagonist1427844_a_at CCAAT/enhancer-binding protein (C/

EBP), beta1431591_s_at Hypothetical protein LOC6771681437270_a_at Cardiotrophin-like cytokine factor 11438364_x_at Angiogenin, member 41440173_x_at Selectin, platelet1443906_at CD55 antigen1448881_at Haptoglobin1449025_at Interferon-induced protein with

tetratricopeptide repeats 31449195_s_at Chemokine (C–X–C motif) ligand 161449254_at Secreted phosphoprotein 11449875_s_at Histocompatibility 2, T region locus 221450297_at Interleukin 61450783_at Interferon-induced protein with

tetratricopeptide repeats 11450826_a_at Serum amyloid A 31451683_x_at Similar to H-2 class I histocompatibility

antigen, L-D alpha precursor1451798_at Interleukin 1 receptor antagonist

AngiogenesisAngiogenesis1421791_at Angiogenin, ribonuclease A family,

member 31438855_x_at Tumor necrosis factor, alpha-induced

protein 21448649_at Glutamyl aminopeptidase1455393_at Ceruloplasmin

Regulation of angiogenesis1448831_at Angiopoietin 21450377_at Similar to thrombospondin 11460302_at Thrombospondin 11448793_a_at Syndecan 41427747_a_at Lipocalin 2

Patterning of blood vessel1434776_at Semaphorin 5A1442340_x_at Cysteine-rich protein 61


both sets of experiments. A list of the twofold changes wascompiled and exported into EASE for further analysis. Thesegroupings were restricted to the GO Biological Process andsubjected to Fisher Exact Probability and Benjamini as theprimary statistics and multiplicity correction respectively.These groupings had to have at least 10 members/functionalgroupings that were part of the GO Biological Process system(p<0.005) to be considered relevant. The relevant lists wereexported back into the affymetrix.com website to compile thegene name and functions.

http://affymetrix.com


Real-time polymerase chain reaction (RT-PCR)

RT-PCR was performed essentially as previously described[31,36]. Cells were cultured until 80% confluence, and serumstarved for 24 h. Total RNA was isolated using Trizol(Invitrogen), and the integrity of the RNA was verified by gelelectrophoresis or Agilent bioanalyzer. Total RNA (25 ng) wasreverse transcribed and amplified using TaqMan Assays onDemand (Applied Biosystems) in a 15-μl reaction volumecontaining two unlabeled primers and 6-carboxyfluorosceinlabeled TaqMan MGB probe. Samples were combined withTaqMan one-step mastermix (Applied Biosystems). Amplifiedsequences were detected using the ABI Prism 7900 HT se-quence detector (Perkin-Elmer-Cetus, Vaudreuil, QC) accord-ing to the manufacturer's instructions. Triplicate sampleswere run, and expression values were standardized to valuesobtained with control 28 S RNA primers using the delta ctmethod. Expression in ccn2+/+ MEFs was taken to be 1.Statistical analysis was performed by the Student's paired ttest. For rescue experiments, cells were serum starved for 18 hand treated for an additional 6 h with or without recombinantCCN2 (Peprotech, Rocky Hill, NJ, 100 ng/ml) prior to harvestingRNA samples for RT-PCR analysis.

Migration assay

Cells were cultured until confluence, and incubated overnightin DMEM, 0.5% FBS. Cells were then incubated in the presenceor absence of 10 μM MEK inhibitor U0126, 100 nM wortmannin(Calbiochem, La Jolla, CA), soluble heparin (10 μg/ml; Sigma),IgG control or anti-syndecan 4 antibody (N-19, which recog-nizes the ectodomain of syndecan 4 and has previously beenshown to block the action of syndecan 4) (Santa Cruz, SantaCruz, CA; [29,30]) for 1 h. Wounding was performed by using apipette tip to make a uniform scratch along the tissue cultureplate. Migration was monitored using a Nikon TE 300 micro-scope. Measurements of distancemigratedwere taken (Image-master, Photon Technology, London, ON). Average ±standarddeviationwas calculated. Statistical analysiswasperformedbythe Student's paired t test.

Cell transfections

Transfections of MEFs were performed as previously described[26–28]. Briefly 2×105 cells were seeded into each well of a 6-well plate the day before transfecting cells with FuGene(Roche, Indianapolis, IN) in a ratio of 3 μl FuGene:2 μg DNAexpression vector. Expression vectors used were: CMV-synde-can 4 (courtesy of Paul Goetinck, Harvard Medical School,Charleston, MA), CMV-dominant negative ras or emptyexpression vector (Upstate). RNAwas harvested and subjectedto real-time PCR analysis as described above.

Immunofluorescence analysis

Cells were permitted to adhere to 24-h glass chamber slides(Nalge Nunc International, Naperville, IL). To detect α-smoothmuscle actin (SMA) or vinculin, cells were fixed for 10 min in3% paraformaldehyde/PBS, blocked in 10% normal goat serum(Jackson ImmunoResearch Laboratories), and subjected to

staining with either mouse anti-SMA antibody (1:100; Sigma-Aldrich) or rabbit anti-vinculin (1:1000; Cell Signaling Tech-nology) antibody followed by fluorescein isothiocyanate(FITC)-conjugated anti-mouse (1:100; Jackson Immuno-Research Laboratories) and Texas Red-conjugated anti-rabbitantibody (1:100; Jackson ImmunoResearch Laboratories),respectively. Cells were photographed using a Zeiss micro-scope and Northern Eclipse (Empix, Missassauga, ON). Imageswere subsequently exported into Adobe Photoshop version 7.0(Adobe Systems, San Jose, CA).

Results

CCN2 is required for the expression of genes involved withtissue remodeling and repair

To evaluate up to what extent CCN2 was required for geneexpression in MEFs, we cultured ccn2+/+ and ccn2−/− MEFsuntil 80% confluence. Cells were incubated for an additional24 h in the absence of serum to remove the confoundinginfluence of other cytokines and growth factors. Total RNAwas prepared from these cells, reverse transcribed, andapplied to Affymetrix MOE430 arrays. Experiments wereperformed twice, and average induction values were obtained.Genespring analysis of data revealed that expression of 474transcripts was reduced >twofold in ccn2+/+ fibroblasts. Genecluster analysis revealed that transcripts involved with tissueremodeling and repair were significantly affected by the lossof CCN2 (Table 1). The overexpressed categories were genesinvolved with cell communication, the immune response, andangiogenesis. Categories were also further subdivided, accord-ing to GO criteria (Table 1). Transcripts reduced in ccn2−/−MEFs included pro-inflammatory chemokines such as CXCL5and interleukin-6 (IL-6) [37–39], angiogenesis- and cancer-associated genes such as ceruloplasmin and lipocalin-2[40,41], and genes promoting ECM remodeling such assyndecan 4, thrombospondin-1 (TSP-1) and aggrecan[35,42,43]. Results were confirmed by real-time polymerasechain reaction analysis (RT-PCR; Fig. 1). Extending ourobservations, proteomic analysis using a multiplex beadassay and a specific IL-6 ELISA showed that protein expressionof the pro-inflammatory and pro-angiogenic cytokines IL-6,CXCL9, fas ligand and angiogenin protein [38,44,45] wasreduced in the absence of CCN2 (Fig. 2). However, reductionof the related IL-1β, IP-10 (CXCL10) and VEGF was notobserved, emphasizing the selectivity of CCN2-dependentaction. Confirming our gene array and RT-PCR results,Western blot analysis showed that syndecan 4 expressionwas reduced in ccn−/− MEFs (Fig. 2).

Our impression that CCN2 was required for the expressionof a subset of genes involved with tissue remodeling andrepair was further evaluated by additional proteomic profiling.A Kinexus KPKS1.2A proteomic screen of ccn2+/+ and ccn2−/−MEFs revealed that ccn2−/− MEFs showed decreased expres-sion of protein kinase D (PKD), protein kinase C ε (PKCε),protein kinase C δ (PKCδ), p21 associated kinase 3 (PAK3),mitogen activated protein kinase kinase 6 (MKK6) andribosomal S6 kinase a (S6Ka) (Fig. 3), which play multiple keyroles, including survival, immune and adhesive/migratory

Fig. 1 – Identification of CCN2-dependent targets in MEFs. mRNA analysis. MEFs isolated from ccn2+/+ and ccn2−/− mice werecultured in DMEM 0.5% FBS for 24 h. RT-PCR was used to detect aggrecan, CXCL5, interleukin-6 (IL-6), thrombospondin-1(TSP-1), syndecan 4, ceruloplasmin and lipocalin-2 mRNA. Basal expression of these transcripts was impaired in ccn2−/−MEFs(p<0.05).


responses [46–50]. Collectively, these results suggested that, inMEFs, CCN2 is required for the expression of mediators oftissue remodeling and repair.

Syndecan 4 is a downstream mediator of CCN2

After we had identified genes whose expression in MEFs wasreduced in the absence of CCN2, we then decided toinvestigate the molecular mechanism underlying the regula-tion of these target genes. Previously, we had demonstrated,using syndecan 4+/+ and syndecan 4−/− adult dermal fibroblasts,that syndecan 4 is required for ERK activation in response toTGFβ and for the elevated ERK activation characterizingfibrotic fibroblasts [35]. Relative to control wild-type MEFs,ccn2−/− MEFs show reduced ERK activation [15]. Compared toIgG control, neutralizing anti-syndecan 4 antibody reduced

ERK phosphorylation in ccn2+/+ MEFs (Fig. 4) confirming ourprior results that syndecan 4 was required for maximal ERKactivation in adult dermal fibroblasts [35]. Given the connec-tion between syndecan 4 and MEK/ERK activation, weassessed whether the characteristics of ccn2−/− MEFs couldbe at least partially attributable to reduced syndecan 4expression. To address this issue, we decided to investigatewhether the reduced expression of mRNAs in ccn2−/− MEFscould be rescued by overexpression of syndecan 4. Given thelink between syndecan 4 and ERK activation in normalfibroblasts, we reasoned that expression of syndecan-4-rescuable transcripts in ccn2−/− MEFs would be MEK/ERK-dependent in ccn2+/+ MEFs. Thus, we returned to our RT-PCRassays and initially screened for genes whose expression inccn2+/+ fibroblasts was MEK/ERK-dependent. We found thatincubation of ccn2+/+ MEFs with the MEK inhibitor U0126

Fig. 2 – Identification of CCN2-dependent targets in MEFs. Protein analysis. MEFs isolated from ccn2+/+ and ccn2−/− micedermal fibroblasts were cultured in DMEM 0.5% FBS for 24 h. (A) CCN2 is required for syndecan 4 expression. Cell lysates wereprobed with antibodies detecting syndecan 4 and GAPDH (HS=heparan sulfated syndecan 4; C=core protein). (B) CCN2 isrequired for the expression of a subset of pro-inflammatory and pro-angiogenic proteins. Equal amounts of conditionedmediafrom ccn2+/+ and ccn2−/− MEFs were subjected to specific ELISAs detecting IL-1beta, IL-6, VEGF, angiogenin, IL-9, IP-10, fasligand and CXCL9. Basal expression of IL-6, angiogenin, IL-9, fas ligand and CXCL9 was reduced in ccn2−/− MEFs (p<0.05).


resulted in a decrease in the expression of the CCN2-dependent mRNAs IL-6 and lipocalin-2 (Fig. 5A). Severalother CCN2-dependent genes, including thrombospondin-1(TSP-1), were MEK-independent suggesting that pathwaysother than MEK/ERK are also involved in the expression ofCCN2-dependent transcripts in ccn2+/+ MEFs (Fig. 5A). Over-expression of dominant negative ras reduced expression ofthe MEK-dependent transcripts IL-6 and lipocalin-2 in ccn2+/+MEFs, but not tsp-1, confirming the involvement of ras/MEK/ERK in the basal expression of these mRNAs in wild-type cells(Fig. 5B).

After having identified a group of transcripts whoseexpression in ccn2+/+ MEFs was MEK/ERK-dependent, it

Fig. 3 – Identification of CCN2-dependent targets in MEFs. Protedermal fibroblasts were cultured in DMEM 0.5% FBS for 24 h andantibodies to the indicated proteins (CDK4, PKD, PKCε, PKCδ, PAProteins were quantified using chemiluminescence and relativemethods). This analysis revealed that ccn2−/−MEFs showed decr(PKCε), protein kinase C δ (PKCδ), PAK3, MKK6 and S6Ka. Lettersindividual proteins listed on the left.

was assessed whether overexpression of syndecan 4 inccn2−/− MEFs might rescue the reduced expression of IL-6and lipocalin-2 mRNA observed in ccn2−/− MEFs. Indeed, wefound that, compared to empty expression vector, transienttransfection of an expression vector encoding syndecan 4into ccn2−/− MEFs resulted in increased IL-6 and lipocalin-2mRNA expression (Fig. 6). These results suggested thatsyndecan 4 is a downstream mediator of at least some ofthe effects of CCN2 in MEFs and differences in phenotypebetween ccn2+/+ and ccn2−/− MEFs are, at least partially anddirectly attributable to the reduced expression of syndecan 4in ccn2−/− MEFs. Addition of recombinant CCN2, containingonly the carboxy-terminal domain IV [4,14] to ccn2−/−

omic analysis. MEFs isolated from ccn2+/+ and ccn2−/− micesubjected to the Kinexus protein kinase screen in whichK3, MKK6, S6Kα, GSK3α, GSKβ, RSK1 and ERK2) are used.expression was determined by Kinexus (see Materials andeased expression of protein kinase D (PKD), protein kinase C εabove individual bands on the protein gel correspond to

Fig. 4 – Syndecan 4 is required for the elevated ERK activationobserved in ccn2+/+ MEFs. Cells (ccn2+/+ or ccn2−/−) wereincubated overnightwith anti-syndecan 4 antibody (anti-syn,2 μg/ml) or IgG control. Cell layers were subjected toWesternblot analyses with anti-total ERK and anti-phospho-ERKantibodies. Anti-syndecan 4 antibody reduced the elevatedERK phosphorylation observed in ccn2+/+ MEFs, yet did notaffect the reduced ERK phosphorylation in ccn2−/− MEFs.


fibroblasts rescued the reduced expression of syndecan 4, IL-6 and lipocalin-2 in this cell type, indicating that thesemRNAs are bone fide targets of CCN2 (Fig. 7).

Fig. 5 – Expression of subset of CCN2-dependent transcripts in ctranscripts whose abundance was reduced by U0126. MEFs isolatthe presence or absence of 10 μM U0126 for 24 h. RT-PCR was usthrombospondin-1 (TSP-1), ceruloplasmin and lipocalin-2 mRNAU0126 (p<0.05). (B) Identification of transcripts whose abundancetransfected with empty expression vector (EMPTY) or expressionhours later, RNAs were harvested and subjected to real-time PCRreduced the expression of the U0126-inhibitable transcripts IL-6 areduced by U0126 treatment.

CCN2 is required for migration of MEFs

A shared feature of morphogenesis in development andtissue repair in adults is elevated fibroblast migration.Syndecan 4 is required for the migration of adult dermalfibroblasts [51]. Thus to provide a functional context for ourgene expression studies in which we showed that syndecan 4was a downstream mediator of CCN2 action, we investigatedwhether (a) CCN2 was required for migration of MEFs and (b)whether this activity was attributable to syndecan 4.Accordingly, we used a scratch wound assay to comparethe migratory abilities of ccn2+/+ and ccn2−/− MEFs. Cellswere grown to confluence and serum starved overnight. Apipette tip was used to make a linear scrape wound throughthe monolayer. Cell migration was monitored for 20 h. Wefound that ccn2+/+ cell MEFs migrated significantly fasterthan ccn2−/− MEFs, indicating that CCN2 was required formaximal migration of MEFs (Fig. 8). To address whether themigration of ccn2+/+ MEFs depended on the HSPG syndecan4, we investigated whether soluble heparin, which blocks

cn2+/+ MEFs is ras/MEK-dependent. (A) Identification ofed from ccn2+/+ mice were cultured in DMEM 0.5% FBS for ined to detect aggrecan, CXCL5, interleukin-6 (IL-6),. Basal expression of IL-6 and lipocalin-2 was reduced bywas reduced by dominant negative ras.Wild-typeMEFswerevector encoding dominant negative ras (dnras). Twenty-fouranalysis. Overexpression of dominant negative ras (dnras)

nd lipocalin-2 but not the transcript TSP-1whichwas also not

Fig. 6 – Reduced expression of lipocalin-2 and IL-6 mRNA in ccn2−/− MEFs is rescued by transfection of expression vectorencoding syndecan 4. ccn2−/− MEFs were transfected with empty expression vector (empty) or expression vector encodingsyndecan 4 (syndecan 4). Twenty-four hours later, RNAs were harvested and subjected to RT-PCR analysis with primersdetecting syndecan 4, lipocalin-2 or IL-6. Average expression ±standard deviation is shown (N=3).


action of heparan sulfate side chains on HSPGs, or aneutralizing anti-syndecan 4 antibody [29,30] could reducethe migration of ccn2+/+ MEFs. Furthermore, as syndecan 4mediates ERK activation in fibroblasts, we assessed whetherthe MEK inhibitor U0126 impaired the migration of ccn2+/+MEFs. We found that either soluble heparin, anti-syndecan 4antibody or U0126 significantly and potently reduced themigration of ccn2+/+ MEF to a level comparable to that ofccn2−/− MEFs, indicating that migration of ccn2+/+ MEFsrequired syndecan 4 and MEK/ERK (Fig. 8).

CCN2-deficient MEFs show defects in focal adhesions, actinfiber formation and ECM contraction

As cell migration requires cycles of actin depolymerizationand polymerization at focal adhesions and focal adhesionturnover [52,53], we investigated whether ccn2−/− MEFs showalso defects in focal adhesions. Thus, ccn2+/+ and ccn2−/−MEFs were cultured for 24 h on glass coverslips, fixed andstained with antibodies detecting α-smooth muscle actin(α-SMA) to visualize actin organization and vinculin toexamine the appearance of focal adhesions (Fig. 9). Asexpected for migratory cells, ccn2+/+ MEFs possessed vinculinat the cell periphery, and actin as organized, directionalbundles of stress fibers. Conversely, ccn2−/− MEFs showed

Fig. 7 – Reduced expression of lipocalin-2 and IL-6 mRNA inccn2−/−MEFs is rescued by recombinant CCN2. ccn2−/−MEFswere treated for 6 h with or without recombinant CCN2(containing only the carboxy-terminal domain). RNAs wereharvested and subjected to RT-PCR analysis with primersdetecting syndecan 4, lipocalin-2 or IL-6. Average expression±standard deviation is shown (N=3).

significantly reduced focal adhesions, and disorganized actinfilaments, consistent with a reduced migratory phenotype(Fig. 9). Actin stress fibers and focal adhesions are required forECM contraction, a key feature of connective tissue remodel-ing. We compared the behavior of ccn2+/+ and ccn2−/− MEFsusing a FPCL assay, whichmeasures the ability of fibroblasts tocontract matrix [31,54], and found that the contractile forcesgenerated across the collagen gel were reduced in ccn2−/−fibroblasts both in the presence and absence of added TGFβ(Fig. 10). These results are consistent with the notion thatCCN2 is required for focal adhesion and actin fiber formation,key components of connective tissue remodeling.

Discussion

In this report, to gain insights into the phenotype observed inccn2−/− mice, we perform the first genome-wide mRNAexpression and proteomic profiling of embryonic fibroblastsgenetically deficient in CCN2. We found that a cohort of genesinvolved with tissue remodeling and repair, including pro-inflammatory, pro-angiogenic and ECM-associated genes wasCCN2-dependent. One gene whose expression was CCN2-dependent was syndecan 4, a protein needed for woundclosure, fibroblast migration and ERK activation in adultfibroblasts [35,51]. Furthermore, we show for the first timethat CCN2-dependent activities in MEFs are, at least partially,due to the expression of the HSPG syndecan 4. Transientoverexpression of syndecan 4 alone restored the reducedexpression of lipocalin-2 and IL-6 mRNAs in ccn2−/− MEFs. Wealso showed for the first time that CCN2 was required for themigration of MEFs, as ccn2−/− MEFs migrated significantlyslower than ccn2+/+ MEFs. Neutralizing anti-syndecan 4 anti-body reduced the migration of ccn2+/+ MEFs to that of ccn2−/−MEFs confirming that syndecan 4 was responsible for theenhanced migration of ccn2+/+ MEFs. Collectively, theseresults indicate that syndecan 4was responsible for conferringaspects of CCN2-dependent action to MEFs. Our resultsshowing that ccn2−/− MEFs migrated significantly slowerthan ccn2+/+ MEFs suggested that CCN2 contributed to theparticipation of fibroblasts to tissue remodeling. Consistentwith this notion, ccn2−/− MEFs showed significantly reducedfocal adhesions, stress fiber formation and a diminishedcapability to generate contractile forces across a collagen gel

Fig. 8 – Migration of MEFs requires CCN2. Migration of ccn2+/+ and ccn2−/− MEFs was monitored, in serum-free media, over20 h using a scratch wound assay. Migration of ccn2−/− MEFs was significantly reduced relative to ccn2+/+ MEFs (Student'spaired t test, p<0.05). Cells were incubated in the presence or absence of U0126 (10 μM), wortmannin (100 nM) and heparinsulfate (10 μg/ml). In addition ccn2+/+ cells were incubated in the presence of 2 μg/ml anti-syndecan 4 antibody or IgG control.Migration of cells over 20 h was measured (N=3) and average migration ±standard deviation is shown. Addition of U0126,heparin or anti-syndecan 4 antibody significantly reduced migration of ccn2+/+ MEFs (Student's paired t test, p<0.05)compared to the appropriate control.


matrix. These results collectively suggest, for the first time,that CCN2 is required forMEFs to actively participate in certaincrucial aspects of tissue remodeling during embryogenesis.

Overexpression of syndecan 4 rescued lipocalin-2 and IL-6expression in ccn2−/− MEFs, in the absence of endogenousCCN2 expression. Thus, in this context, syndecan 4 is notacting as a CCN2 receptor, as syndecan 4 is able to exert itseffects independently of CCN2, presumably through its ability

to interact with other growth factors and/or ECM [55]. Theseresults, along with the observation that neutralizing syndecan4 antibody reduced the migration of ccn2+/+ MEFs to that ofccn2−/− MEFs, confirm that CCN2-dependent action in wild-type MEFs relies on syndecan 4 expression and that thephenotype of ccn2−/− MEFs is directly attributable, at least inpart, to a loss of syndecan 4 expression. Please note thatwhether syndecan 4 can function directly as a CCN2 receptor

Fig. 9 – CCN2 is required for focal adhesion and actin stress fiber formation MEFs were fixed in paraformaldehyde andincubated with anti-α-SMA or anti-vinculin antibodies, followed by appropriate fluorescently tagged secondary antibodies.Whereas ccn2+/+ fibroblasts show peripheral localization of focal adhesions (as visualized by anti-vinculin staining) and adirectional orientation of stress fibers consistent withmotile cells, ccn2−/−MEFs show a disorganized actin network consistentwith cells whose migration is impaired.


to mediate CCN2 activity is beyond the scope of the currentstudy; however, CCN2 is capable of activating ERK in syndecan4−/− adult dermal fibroblasts (Y.C. and A.L., unpublished data).That said, our results described in this report clearly suggestthat syndecan 4 acts downstream of CCN2, but independent ofa direct involvement of CCN2 protein. However, as previouslywe have shown that in ccn2+/+ MEFs CCN2 interacts withsyndecan 4, a receptor for fibronectin, and that ccn2−/− MEFs

Fig. 10 – CCN2 is required for contractile forces acrosscollagen gel matrices. The effect of loss of CCN2 expressionon contractile force generated by MEFs in a fixed, tetheredfloating collagen gel lattice was investigated using a CultureForce Monitor. Note the higher contractile forces generatedby cells derived from ccn2+/+ (WT) in comparison toccn2−/− (KO) fibroblasts. Traces are the result of using 5 ml ofcollagen gel at a density of 1 million cells per milliliter ofcollagen gel. A representative trace is shown (N=3).

are not able to properly spread on fibronectin, syndecan 4might act as a receptor for CCN2 depending on the context [15].It is interesting to note that syndecan 4 knockout mice areviable, showing modest defects in tissue repair, fibroblastmigration and angiogenesis [51]. These results are in contrastto CCN2 knockout mice which show a more severe perinatallethal phenotype [25]. Thus CCN2 is likely to also act throughsyndecan-4-independent pathways. Identification of thesesyndecan-4-independent mechanisms is underway, but isbeyond the scope of this current study.

Although other CCN family members have similar func-tions in vitro [22–24,56], ccn2−/− and ccn1−/− show distinctphenotypes in vivo [25,57], illustrating that CCN1 and CCN2perform essential, physiologically non-redundant functions invivo. Consistent with this notion, our report is the first tosuggest that CCN2 provides essential non-redundant func-tions to MEFs. Specifically, CCN2 is required for gene expres-sion, migration and contraction of ECM as well as focaladhesion and stress fiber formation. Our results furtherindicate that CCN2 plays a key role in ECM remodeling byMEFs. As the tissue repair and fibrotic program in adults isbelieved to be a recapitulation of the matrix remodelingprocess in embryonic development [58], it is likely that CCN2plays a key role in promoting tissue formation, regenerationand repair in adults. CCN proteins, including CCN2, can oftenact by modifying the action of other proteins [4,59,60]. In thisreport we found that a form of recombinant CCN2, containingthe carboxy-terminal domain IV, was able to cause a modestyet significant induction of syndecan 4, IL-6 and lipocalin-2gene expression supporting the notion that CCN proteins can


have some independent activity [4.60]. That re-introduction ofsyndecan 4 into the ccn2−/− MEFs resulted in the induction ofsome genes whose expression was decreased in ccn2−/− MEFsemphasizes a central conclusion of our paper: that thephenotype of ccn2−/− MEFs is at least partially attributable toa reduction in syndecan 4 expression.

In summary, our results provide new insights into CCN2function in fibroblasts. Our data reveal that syndecan 4 actsdownstream of CCN2 and that loss of syndecan 4 expression isresponsible for at least part of the ccn2−/− phenotype. Further,our results suggesting that CCN2 is required for the migrationand contraction of MEFs support the idea that CCN2 plays acritical non-redundant role in facilitating tissue remodelingand repair. Our results are consistent with the notion thatanti-CCN2 strategies may selectively and significantly alterthe cell–matrix interactions that characterize the excessivetissue repair programs hypothesized to result in connectivetissue pathologies including fibrosis and cancer [61,62].

Acknowledgments

This work was supported in part by The Arthritis ResearchCampaign, The Raynaud's and Scleroderma Association Trust,theWelton Foundation, the Scleroderma Society, the NationalInstitute of Health, the Canadian Foundation for Innovationand the Canadian Institutes of Health Research. A.L. is anArthritis Society (Scleroderma Society of Ontario) New Inves-tigator and an Early Investigator Award recipient. We thankKelly Summers for expert technical assistance in performingbead assays.

R E F E R E N C E S

[1] S.F. Badylak, The extracellular matrix as a scaffold for tissuereconstruction, Semin. Cell Dev. Biol. 13 (2002) 377–383.

[2] L.F. Lau, S.C. Lam, The CCN family of angiogenic regulators:the integrin connection, Exp. Cell Res. 248 (1999) 44–57.

[3] E.E. Moussad, D.R. Brigstock, Connective tissue growth factor:what's in a name? Mol. Genet. Metab. 71 (2000) 276–292.

[4] B. Perbal, CCN proteins: multifunctional signalling regulators,Lancet 363 (2004) 62–64.

[5] A. Leask, D.J. Abraham, The role of connective tissue growthfactor, a multifunctional matricellular protein, in fibroblastbiology, Biochem. Cell Biol. 81 (2003) 355–363.

[6] A. Igarashi, H. Okochi, D.M. Bradham, G.R. Grotendorst,Regulation of connective tissue growth factor geneexpression in human skin fibroblasts and during woundrepair, Mol. Biol. Cell 4 (1993) 637–645.

[7] G.A. Surveyor, A.K. Wilson, D.R. Brigstock, Localization ofconnective tissue growth factor during the period of embryoimplantation in the mouse, Biol. Reprod. 59 (1998)1207–1213.

[8] S. Friedrichsen, H. Heuer, S. Christ, M. Winckler, D. Brauer, K.Bauer, G. Raivich, CTGF expression during mouse embryonicdevelopment, Cell Tissue Res. 312 (2003) 175–188.

[9] K.S. Frazier, G.R. Grotendorst, Expression of connective tissuegrowth factor mRNA in the fibrous stroma of mammarytumors, Int. J. Biochem. Cell Biol. 29 (1997) 153–161.

[10] D.J. Abraham, X. Shiwen, C.M. Black, S. Sa, Y. Xu, A. Leask,Tumor necrosis factor alpha suppresses the induction ofconnective tissue growth factor by transforming growth

factor-beta in normal and scleroderma fibroblasts, J. Biol.Chem. 275 (2000) 15220–15225.

[11] I.E. Blom, R. Goldschmeding, A. Leask, Gene regulation ofconnective tissue growth factor: new targets for antifibrotictherapy? Matrix Biol. 21 (2002) 473–482.

[12] A.M. Babic, C.C. Chen, L.F. Lau, Fisp12/mouse connectivetissue growth factor mediates endothelial cell adhesion andmigration through integrin alphavbeta3, promotesendothelial cell survival, and induces angiogenesis in vivo,Mol. Cell. Biol. 19 (1999) 2958–2966.

[13] C.C. Chen, N. Chen, L.F. Lau, The angiogenic factors Cyr61 andconnective tissue growth factor induce adhesive signaling inprimary human skin fibroblasts, J. Biol. Chem. 276 (2001)10443–10452.

[14] R. Gao, D.R. Brigstock, Connective tissue growth factor (CCN2)induces adhesion of rat activated hepatic stellate cells bybinding of its C-terminal domain to integrin alpha(v)beta(3)and heparan sulfate proteoglycan, J. Biol. Chem. 279 (2004)8848–8855.

[15] Y. Chen, D.J. Abraham, X. Shi-Wen, J.D. Pearson, C.M. Black,K.M. Lyons, A. Leask, CCN2 (connective tissue growth factor)promotes fibroblast adhesion to fibronectin, Mol. Biol. Cell 15(2004) 5635–5646.

[16] M.L. Kireeva, B.V. Latinkic, T.V. Kolesnikova, C.C. Chen, G.P.Yang, A.S. Abler, L.F. Lau, Cyr61 and Fisp12 are bothECM-associated signaling molecules: activities, metabolism,and localization during development, Exp. Cell Res. 233 (1997)63–77.

[17] E. Gore-Hyer, J. Pannu, E.A. Smith, G. Grotendorst, M.Trojanowska, Selective stimulation of collagen synthesis inthe presence of costimulatory insulin signaling by connectivetissue growth factor in scleroderma fibroblasts, ArthritisRheum. 48 (2003) 798–806.

[18] G. Hashimoto, I. Inoki, Y. Fujii, T. Aoki, E. Ikeda, Y. Okada,Matrix metalloproteinases cleave connective tissue growthfactor and reactivate angiogenic activity of vascularendothelial growth factor 165, J. Biol. Chem. 277 (2002)36288–36295.

[19] G.R. Grotendorst, H. Rahmanie, M.R. Duncan, Combinatorialsignaling pathways determine fibroblast proliferation andmyofibroblast differentiation, FASEB J. 18 (2004) 469–479.

[20] G.R. Grotendorst, M.R. Duncan, Individual domains ofconnective tissue growth factor regulate fibroblastproliferation and myofibroblast differentiation, FASEB J. 19(2005) 729–738.

[21] X. Shi-wen, L.A. Stanton, L. Kennedy, D. Pala, Y. Chen, S.L.Howat, E.A. Renzoni, D.E. Carter, G. Bou-Gharios, R.J. Stratton,J.D. Pearson, F. Beier, K.M. Lyons, C.M. Black, D.J. Abraham,A. Leask, CCN2 is necessary for adhesive responses totransforming growth factor-beta1 in embryonic fibroblasts,J. Biol. Chem. 281 (2006) 10715–10726.

[22] M.L. Kireeva, S.C. Lam, L.F. Lau, Adhesion of human umbilicalvein endothelial cells to the immediate-early gene productCyr61 ismediated through integrin alphavbeta3, J. Biol. Chem.273 (1998) 3090–3096.

[23] C.G. Lin, S.J. Leu, N. Chen, C.M. Tebeau, S.X. Lin, C.Y. Yeung,L.F. Lau, CCN3 (NOV) is a novel angiogenic regulator of theCCN protein family, J. Biol. Chem. 278 (2003) 24200–24208.

[24] J. Lafont, H. Thibout, C. Dubois, M. Laurent, C. Martinerie,NOV/CCN3 induces adhesion of muscle skeletal cells andcooperates with FGF2 and IGF-1 to promote proliferation andsurvival, Cell Commun. Adhes. 12 (2005) 41–57.

[25] S. Ivkovic, B.S. Yoon, S.N. Popoff, F.F. Safadi, D.E. Libuda, R.C.Stephenson, A. Daluiski, K.M. Lyons, Connective tissuegrowth factor coordinates chondrogenesis and angiogenesisduring skeletal development, Development 130 (2003)2779–2791.

[26] A. Holmes, D.J. Abraham, S. Sa, X. Shiwen, C.M. Black, A.Leask, CCN2 and SMADs, maintenance of scleroderma


phenotype is independent of SMAD signaling, J. Biol. Chem.276 (2001) 10594–10601.

[27] A. Leask, A. Holmes, C.M. Black, D.J. Abraham, Connectivetissue growth factor gene regulation. Requirements for itsinduction by transforming growth factor-beta 2 in fibroblasts,J. Biol. Chem. 278 (2003) 13008–13015.

[28] J.P. Van Beek, L. Kennedy, J.S. Rockel, S.M. Bernier, A. Leask,The induction of CCN2 by TGFbeta1 involves Ets-1, ArthritisRes. Ther. 8 (2006) R36.

[29] L. Lin, M. Kurpakus-Wheater, Laminin alpha5 chain adhesionand signaling in conjunctival epithelial cells, Invest.Ophthalmol. Visual Sci. 43 (2002) 2615–2621.

[30] C. Feistritzer, N.C. Kaneider, D.H. Sturn, C.J. Wiedermann,Syndecan-4-dependent migration of human eosinophils,Clin. Exp. Allergy 34 (2004) 696–703.

[31] M. Eastwood, R. Porter, U. Khan, G. McGrouther, R. Brown,Quantitative analysis of collagen gel contractile forcesgenerated by dermal fibroblasts and the relationship to cellmorphology, J. Cell. Physiol. 166 (1996) 33–42.

[32] F. Grinnell, C.H. Ho, Transforming growth factor betastimulates fibroblast-collagenmatrix contraction by differentmechanisms in mechanically loaded and unloaded matrices,Exp. Cell Res. 273 (2002) 248–255.

[33] X. Shi-Wen, Y. Chen, C.P. Denton, M. Eastwood, E.A. Renzoni,G. Bou-Gharios, J.D. Pearson, M. Dashwood, R.M. Du Bois, C.M.Black, A. Leask, D.J. Abraham, Endothelin-1 promotesmyofibroblast induction through the ETA receptor via a rac/phosphoinositide 3-kinase/Akt-dependent pathway and isessential for the enhanced contractile phenotype of fibroticfibroblasts, Mol. Biol. Cell 15 (2004) 2707–2719.

[34] R.A. Brown, R. Prajapati, D.A. McGrouther, I.V. Yannas, M.Eastwood, Tensional homeostasis in dermal fibroblasts:mechanical responses to mechanical loading inthree-dimensional substrates, J. Cell. Physiol. 175 (1998)323–332.

[35] Y. Chen, X. Shi-Wen, J. van Beek, L. Kennedy, M. McLeod, E.A.Renzoni, G. Bou-Gharios, S. Wilcox-Adelman, P.F. Goetinck,M. Eastwood, C.M. Black, D.J. Abraham, A. Leask, Matrixcontraction by dermal fibroblasts requires transforminggrowth factor-beta/activin-linked kinase 5, heparansulfate-containing proteoglycans, and MEK/ERK: insights intopathological scarring in chronic fibrotic disease, Am. J. Pathol.167 (2005) 1699–1711.

[36] X. Shi-wen, S.L. Howat, E.A. Renzoni, A. Holmes, J.D. Pearson,M.R. Dashwood, G. Bou-Gharios, C.P. Denton, R.M. du Bois,C.M. Black, A. Leask, D.J. Abraham, Endothelin-1 inducesexpression of matrix-associated genes in lung fibroblaststhrough MEK/ERK, J. Biol. Chem. 279 (2004) 23098–23103.

[37] L. Holmdahl, M.L. Ivarsson, The role of cytokines, coagulation,and fibrinolysis in peritoneal tissue repair, Eur. J. Surg. 165(1999) 1012–1019.

[38] L.R. Bisset, P. Schmid-Grendelmeier, Chemokines and theirreceptors in the pathogenesis of allergic asthma: progress andperspective, Curr. Opin. Pulm. Med. 11 (2005) 35–42.

[39] A.R. Gear, D. Camerini, Platelet chemokines and chemokinereceptors: linking hemostasis, inflammation, and hostdefense, Microcirculation 10 (2003) 335–350.

[40] S.A. Lowndes, A.L. Harris, Copper chelation as anantiangiogenic therapy, Oncol. Res. 14 (2004) 529–539.

[41] A. Friedl, S.P. Stoesz, P. Buckley, M.N. Gould, Neutrophilgelatinase-associated lipocalin in normal and neoplastichuman tissues. Cell type-specific pattern of expression,Histochem. J. 31 (1999) 433–441.

[42] H. Chen, M.E. Herndon, J. Lawler, The cell biology ofthrombospondin-1, Matrix Biol. 19 (2000) 597–614.

[43] J.T. Gallagher, The extended family of proteoglycans: socialresidents of the pericellular zone, Curr. Opin. Cell Biol. 1 (1989)1201–1218.

[44] C. Lambert, A.M. Landau, J. Desbarats, Fas-beyond death: aregenerative role for Fas in the nervous system, Apoptosis 8(2003) 551–562.

[45] S. Ugurel, G. Rappl, W. Tilgen, U. Reinhold, Increased serumconcentration of angiogenic factors in malignant melanomapatients correlates with tumor progression and survival,J. Clin. Oncol. 19 (2001) 577–583.

[46] H. Oster, D. Abraham, M. Leitges, Expression of the proteinkinase D (PKD) family during mouse embryogenesis, GeneExpression Patterns 6 (2006) 400–408.

[47] E. Aksoy, M. Goldman, F. Willems, Protein kinase C epsilon: anew target to control inflammation and immune-mediateddisorders, Int. J. Biochem. Cell Biol. 36 (2004) 183–188.

[48] D.N. Jackson, D.A. Foster, The enigmatic protein kinaseCdelta: complex roles in cell proliferation and survival, FASEBJ. 18 (2004) 627–636.

[49] Z.M. Jaffer, J. Chernoff, p21-activated kinases: three more jointhe Pak, Int. J. Biochem. Cell Biol. 34 (2002) 713–717.

[50] L.A. Berven, F.S. Willard, M.F. Crouch, Role of the p70(S6K)pathway in regulating the actin cytoskeleton and cellmigration, Exp. Cell Res. 296 (2004) 183–195.

[51] F. Echtermeyer, M. Streit, S. Wilcox-Adelman, F. Saoncella, M.Denhez, M. Detmar, F.P. Goetinck, Delayed wound repair andimpaired angiogenesis in mice lacking syndecan-4, J. Clin.Invest. 107 (2001) R9–R14.

[52] D.J. Sieg, C.R. Hauck, D.D. Schlaepfer, Required role of focaladhesion kinase (FAK) for integrin-stimulated cell migration,J. Cell Sci. 112 (1999) 2677–2691.

[53] A.J. Ridley, W.E. Allen, M. Peppelenbosch, G.E. Jones, Rhofamily proteins and cell migration, Biochem. Soc. Symp. 65(1999) 111–123.

[54] M. Eastwood, V.C. Mudera, D.A. McGrouther, R.A. Brown, Theeffect of precise mechanical loading on fibroblast populatedcollagen lattices: morphological changes, Cell Motil.Cytoskelet. 49 (1998) 13–21.

[55] M. Simons, A. Horowitz, Syndecan-4-mediated signalling,Cell. Signalling 13 (2001) 855–862.

[56] S. Benini, B. Perbal, D. Zambelli, M.P. Colombo, M.C. Manara,M. Serra, M. Parenza, V. Martinez, P. Picci, K. Scotlandi, InEwing's sarcoma CCN3(NOV) inhibits proliferation whilepromoting migration and invasion of the same cell type,Oncogene 24 (2005) 4349–4361.

[57] F.E. Mo, A.G. Muntean, C.C. Chen, D.B. Stolz, S.C. Watkins,L.F. Lau, CYR61 (CCN1) is essential for placentaldevelopment and vascular integrity, Mol. Cell. Biol. 22(2002) 8709–8720.

[58] P. Martin, S.M. Parkhurst, Parallels between tissue repairand embryo morphogenesis, Development 131 (2004)3021–3034.

[59] D.R. Brigstock, The CCN family: a new stimulus package,J. Endocrinol. 178 (2003) 169–175.

[60] A. Leask, D.J. Abraham, All in the CCN family: essentialmatricellular modulators emerge from the bunker,J. Cell Sci. 119 (2006) 4803–4810.

[61] A. Leask, D.J. Abraham, The control of TGFβ signaling inwound healing and fibrosis, FASEB J. 18 (2004) 816–827.

[62] A. Leask, Scar wars: is TGFbeta the phantom menace inscleroderma? Arthritis Res. Ther. 8 (4) (2006) 213.

ccn2 is necessary for the function of mouse embryonic fibroblasts

Documents