J Biol Chem, Vol. 274, Issue 32, 22476-22483, August 6, 1999

DANCE, a Novel Secreted RGD Protein Expressed in Developing, Atherosclerotic, and Balloon-injured Arteries^*

From the ^a Department of Medical Chemistry, Faculty of Medicine, ^b Cardiovascular Medicine, ^h Center for Molecular Biology and Genetics, Kyoto University, Sakyo-ku, Kyoto 606, Japan, ^c Department of Medicine and Center for Molecular Genetics, University of California at San Diego, La Jolla, California 92093-0613, ^f Maine Medical Center Research Institute, South Portland, Maine 04106, ^g Third Department of Internal Medicine, Kyoto Prefectural University of Medicine, Kamigyo-ku, Kyoto 602, Japan, and ⁱ Genetics Institute, Inc., Cambridge, Massachusetts 02140

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS AND DISCUSSION REFERENCES

We have identified and characterized mouse, rat, and human cDNAs that encode a novel secreted protein of 448 amino acids named DANCE (developmental arteries and neural crest epidermal growth factor (EGF)-like). DANCE contains six calcium-binding EGF-like domains, one of which includes an RGD motif. Overexpression studies of recombinant DANCE protein document that DANCE is a secreted 66-kDa protein. DANCE and recently described protein S1-5 comprise a new EGF-like protein family. The human DANCE gene was mapped at chromosome 14q32.1. DANCE mRNA is mainly expressed in heart, ovary, and colon in adult human tissues. Expression profile analysis by in situ hybridization revealed prominent DANCE expression in developing arteries. DANCE is also expressed in neural crest cell derivatives, endocardial cushion tissue, and several other mesenchymal tissues. In adult vessels, DANCE expression is largely diminished but is reinduced in balloon-injured vessels and atherosclerotic lesions, notably in intimal vascular smooth muscle cells and endothelial cells that lose their ability to proliferate in late stage of injury. DANCE protein was shown to promote adhesion of endothelial cells through interaction of integrins and the RGD motif of DANCE. DANCE is thus a novel vascular ligand for integrin receptors and may play a role in vascular development and remodeling.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS AND DISCUSSION REFERENCES

Vascular development in the growing embryo requires the controlled proliferation of smooth muscle cells, endothelial cells, and fibroblasts and their continuous remodeling to form larger vessels. In most adult tissues, these vascular component cells are normally quiescent. However, vascular remodeling also plays an important role in many cardiovascular disorders (1). For example, in atherosclerosis and in restenosis after balloon angioplasty, vascular smooth muscle cells migrate and proliferate in the intima, resulting in the narrowing of the vascular lumen (2). In these conditions, cells are likely to reactivate fetal programs to enter the cell cycle. Indeed, the proliferating smooth muscle cells in the intima have changed from a contractile phenotype that can respond to vasoconstriction or vasodilation signals to a synthetic phenotype that can respond to growth stimulation. The synthetic phenotype may be regarded as an embryonic phenotype, as indicated by the switch in expression to an embryonic isoform of myosin heavy chain (3).

Extracellular matrix proteins are intimately involved in vascular remodeling, by affecting growth, migration, differentiation, and survival of vascular cell types (4). Integrins constitute a large family of cell surface receptors for extracellular matrix proteins that mediate not only cell adhesion by cell-matrix and cell-cell interaction but also multiple outside-in signals that lead to activation of downstream pathways such as tyrosine kinases and phosphatidylinositol 3-kinase (5). The most common integrin recognition sites of ligands contain a consensus Arg-Gly-Asp (RGD) motif. This RGD motif is recognized by many integrins (₅1, _IIb3, and all v integrins), and synthetic peptides containing this RGD sequence are known to antagonize these integrins and inhibit angiogenesis or thrombosis (6).

Another important motif frequently seen in secreted proteins, including diverse vascular extracellular matrix proteins, is an epidermal growth factor (EGF)¹-like motif (7). A subset of the EGF-like domains contain a distinctive amino acid sequence motif that is associated with calcium binding (8). This calcium-binding EGF-like (cbEGF) motif is found in several extracellular matrix proteins (fibrillin-1 and -2, fibulin-1 and -2, and nidogen) that are distributed in vessel walls, as well as in regulators of blood coagulation (factors IX and X, proteins C and S, and thrombomodulin), low density lipoprotein receptor, transforming growth factor--binding protein, Notch, and its ligands Delta and Serrate, which are known to be involved in cell differentiation. Tandemly repeated cbEGF domains in the presence of calcium ion form structurally stable helices that allow protein-protein interaction by these domains (9).

To isolate novel secreted molecules involved in the control of cardiovascular development and disease, we have employed the signal sequence trap method (10, 11), which is a newly developed cloning technique to isolate cDNAs of proteins containing signal sequences, including secreted, membrane, and endoplasmic reticulum proteins. We screened a cDNA library from mouse embryonic heart and isolated several new molecules. In this report, we describe the cloning and characterization of a novel secreted protein with an RGD motif and cbEGF domains. This molecule, named DANCE (developmental arteries and neural crest EGF-like) protein after its embryonic distribution, is expressed mainly in developmental arteries and is re-expressed in atherosclerotic lesions and in balloon-injured arteries. DANCE promotes cell attachment via interaction of integrins and the RGD motif of the protein. DANCE is thus a new ligand of integrins that may play an important role in vascular remodeling.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS AND DISCUSSION REFERENCES

Signal Sequence Trap Screening, cDNA Cloning, and Northern Blot-- Poly(A) RNA from 9.5- and 18-dpc mouse embryonic heart was extracted with TRIzol reagent (Life Technologies, Inc.) and Oligotex-dT30 Super (Roche Molecular Biochemicals). The construction of the cDNA library and screening by yeast signal sequence trap was carried out as described previously (11). Full-length cDNAs were cloned by screening a mouse 13.5-dpc embryonic heart cDNA library (Stratagene). Two positive clones were sequenced. 5'-Rapid amplification of cDNA ends (RACE) was performed using the Marathon cDNA amplification kit (CLONTECH) according to the manufacturer's protocol, and six clones were sequenced. The rat DANCE homologue clone was obtained by screening a rat heart cDNA library (Stratagene) and by 5'-RACE. The human DANCE clone was obtained from IMAGE (GenBank^TM accession no. H17726). Each clone was sequenced and analyzed with the computer analysis program GeneWorks (IntelliGenetics, Inc.). A homology search was performed with BLAST and FASTA against public sequence data bases, and a motif search was performed on line at Prosite. Multiple sequence alignment was carried out using an on-line program.² For Northern blot analysis, a human multiple tissue Northern blot (CLONTECH) was hybridized to [³²P]dCTP-labeled full-length human DANCE cDNA (2.5 kilobase pairs) using QuickHyb (Stratagene). For Northern analysis of rat carotid arteries, RNA was extracted with TRIzol (Life Technologies, Inc.), and [³²P]dCTP-labeled full-length rat DANCE cDNA (2.5 kilobase pairs) was used as a probe. All blots were washed at 0.1× SSC, 65 °C, and exposed to autoradiography film overnight.

Chromosomal Mapping by Fluorescence in Situ Hybridization-- Full-length human DANCE cDNA was used as a probe. The probe was labeled by standard nick translation using biotin-16-dUTP (Roche Molecular Biochemicals) and was purified using Sephadex G-50 spin columns. Hybridization and detection by fluorescein isothiocyanate was carried out as described previously (12). Chromosomes were identified using counterstaining with 4',6'-diaminido-2-phenolindole dihydrochloride.

Protein Expression and Antibodies-- An EcoRI fragment including the full coding sequence of rat DANCE cDNA was cloned into a pCXN expression vector (gift from Dr. M. Kinoshita at Kyoto University), where expression is driven by the chick -actin promoter and cytomegalovirus enhancer. Transfections of COS7 and 293T cells were performed using Lipofectamine Plus (Life Technologies, Inc.), according to the manufacturer's protocol. 16 h after transfection, the medium was changed to 8 ml of serum-free Dulbecco's modified Eagle's medium per 10-cm plate. 24 h later, the medium was harvested, and cells from 10-cm plates were lysed in 8 ml of radioimmune precipitation assay buffer (1× phosphate-buffered saline, 1% Nonidet P-40, 0.1% SDS). 25 µl of medium and cell lysate were analyzed by SDS-polyacrylamide gel electrophoresis and immunoblotting. Rabbit anti-mouse/rat DANCE polyclonal antibodies were raised against KLH-conjugated polypeptide CMTRPIKGPRDIQLDLEMITVN, which corresponds to amino acids 406-426 of mouse and rat DANCE protein by TANA laboratories, LC. Western blotting was carried out following the ECL Western blot (Amersham Pharmacia Biotech) protocol.

In Situ Hybridization-- For in situ hybridization of mouse tissues, a XbaI-PstI fragment (616 base pairs) was subcloned into pBlueScript in both directions. The sense and antisense RNA probe was transcribed in vitro. 7-µm paraffin sections from either mouse tissues or embryos were rehydrated, treated with 10 mg/ml proteinase K for 7.5 min, and hybridized with 10,000 cpm of ³⁵S-labeled riboprobe/ml in hybridization solution containing 50% formamide, 30 mM NaCl, 20 mM EDTA, 10 mM NaH₂PO₄, 10% dextran sulfate, 1× Denhardt's solution, and 10 mM dithiothreitol. Hybridization was carried out for 14 h at 60 °C. Samples were subsequently washed at 60 °C in 5× SSC, 10 mM dithiothreitol, 2× SSC, 50% formamide, 10 mM dithiothreitol; digested with RNase A (10 mg/ml) for 30 min at 37 °C; and dehydrated. Slides were dipped in Ilford-K5 photographic emulsion (Polysciences, Warrington, PA), exposed for 3 weeks, and developed in Kodak D19 solution. Counterstaining was performed in 0.02% toluidin blue/hematoxylin.

In situ hybridization of rat balloon-injured aortae was carried out on en face preparations of vessel segments as described (13), using [³⁵S]UTP-labeled sense and antisense rat DANCE riboprobes. In this model, endothelial regeneration and smooth muscle cell proliferation were studied in the thoracic aorta of which the endothelium had been removed with a 2F Fogarty balloon catheter. In this vessel, endothelial regrowth occurs from intercostal arteries, and rapidly proliferating smooth muscle cells migrate onto the denuded surface from the underlying tunica media starting at approximately 6-8 days after denudation. Once smooth muscle cells have migrated into the intima, they form an intimal lesion over the course of several weeks. By examining the cells on the luminal surface, this model allows the study of luminal smooth muscle cells at early times after injury, when cells are replicating, and at late times, when proliferation stops. Deendothelialized segments of arteries were identified by intravenous injection of Evans blue (0.3 ml of 5% solution in saline) 10 min prior to sacrifice. All animals were perfusion-fixed with phosphate (0.1 M, pH 7.4)-buffered 4% paraformaldehyde. For en face preparations, vessel segments were cut open longitudinally, and the tissue was pinned out flat on Teflon cards (luminal side facing up). Subsequent steps followed the protocol were as described (13).

Cell Attachment Assay and Attachment Inhibition Assay-- Human DANCE cDNA without the termination codon were subcloned to pEF6/V5 (Invitrogen) to add a C-terminal His₆ tag, and transfected to 293T cells. Recombinant protein was purified from the serum-free conditioned medium of stable lines, using nickel-nitrilotriacetic acid-agarose (Qiagen) according to established protocol, and dialyzed into Hanks' balanced salt solution containing 10 mM HEPES. Protein purity was confirmed by Coomassie Blue staining of SDS-polyacrylamide gel and Western blot. Protein concentration was determined by Coomassie Plus reagent (Pierce) using BSA as a standard.

Enzyme-linked immunosorbent assay 96-well plates (Nunc) were coated with 0.5-16 µg/ml of either DANCE protein or BSA diluted in Hanks' balanced salt solution for 18 h at 4 °C. The plates were blocked for 1 h with a solution of heat-treated phosphate-buffered saline containing 10 mg/ml BSA. Human umbilical vein endothelial cells (HUVECs, from Clonetics) were harvested by trypsinization and resuspended in adhesion buffer (Hanks' balanced salt solution containing 10 mM HEPES, 2.2 mM MgCl₂, 0.2 mM MnCl₂, and 1% BSA). Cells (1.5 × 10⁴/100µl) were added to each well in the presence or absence of peptides. Peptide antagonists included GRGDSP and control peptide GRGESP (both from Life Technologies, Inc.) at concentrations of 25-800 µM were preincubated with cells for 30 min before being placed in each well. Cells were incubated at 37 °C for 90 min, and wells were washed several times. 100 µl of medium was added to each well, and relative cell number was determined with the Cell Titer AQ reagent (Promega).

	RESULTS AND DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS AND DISCUSSION REFERENCES

Cloning and Sequencing of a Novel cDNA A55 (DANCE)-- We screened 6.6 × 10⁶ yeast transformants from mouse 18-dpc embryonic heart and 6.0 × 10⁶ yeast transformants from mouse 9.5-dpc embryonic heart, and 647 positive clones were obtained by the yeast signal sequence trap method, as described previously (14). All positive clones were sequenced, and redundant clones were removed. Of 62 independent clones, 33 were identical or homologous to sequences reported in mice or other mammals. One of the novel clones, A55, contained an EGF-like domain. Because the clones obtained by the screening contained only N-terminal sequence, poly(A)-tailed cDNAs for A55 were isolated from a murine 13.5-dpc embryonic heart cDNA library. 5' cDNAs containing the complete 5'-untranslated region were obtained by 5'-RACE. The sequence encodes a 448-amino acid protein (Fig. 1A). The translation start site methionine was assigned at nucleotide positions 320-322 because of the presence of an upstream stop codon (positions 266-268), N-terminal signal sequence, and the compatibility with a Kozak consensus sequence (15). The deduced amino acid sequence has an N-terminal hydrophobic domain, which is presumed to be the signal sequence and is predicted to be cleaved after Ala at position 23 but has no other hydrophobic regions that can serve as transmembrane domains and does not have a C-terminal endoplasmic reticulum retention signal (KDEL and related sequences), suggesting that A55 is a secreted protein. A55 was named as DANCE (developmental arteries and neural crest EGF-like) to represent its expression profile. Sequencing of the six clones from 5'-RACE revealed that only one contained a different 5'-untranslated region and N-terminal coding sequence (Fig. 1A). By sequencing a genomic clone containing the 5' region of DANCE, this sequence was found to exist just downstream of major exon 1 (data not shown).

View larger version (93K): [in this window] [in a new window]   Fig. 1.   A, nucleotide sequence and deduced amino acid sequence of mouse A55 (DANCE) cDNA. Nucleotide numbers are based on the sequence of the major transcript, and bracketed nucleotide numbers indicate that of the minor, alternative spliced form of the transcript. Predicted signal sequence cleavage site using a program based on von Heijne's data (32) is shown by an arrowhead. An RGD motif found in the first cbEGF domain is boxed. Two putative Asn glycosylation sites are indicated by asterisks. B, alignment of six cbEGF domains of mouse DANCE with that of mouse EGF precursor (EGF), Notch, LDL receptor (LDLR), vitamin K-dependent protein S (PRTS), fibrillin 1 (FBN1), and fibulin-1 (FBLN1). Numbers indicate amino acid positions. Conserved residues including cysteines and those required for calcium binding are shaded. C, alignment of mouse, rat, and human DANCE amino acid sequences. For rat and human DANCE, only amino acid residues differing from those of the mouse clone are shown, and conserved amino acid residues are indicated by dashes. D, schematic diagram comparing DANCE, UPH1 (GenBankTM AF093119)/H411 (AF046870), and S1-5 (U03877)/T16 (D89730). N-terminal signal sequences are indicated as black boxes, and cbEGF domains are shown as rounded rectangles. The RGD motif, which is only found in DANCE, is also indicated. Amino acid identity of the cbEGF repeat and C-terminal domains between human clones of the family are shown below. E, alignment of the C-terminal domains of mouse DANCE and mouse fibulin-1C, fibulin-1D, and fibulin-2. Conserved amino acid residues that are identical to DANCE are shaded. Numbers indicate amino acid positions.

The DANCE transcript encodes a protein with six EGF-like domains (Fig. 1B). One is located at the N terminus, and the other five are tandemly distributed at the center of the molecule. These domains contain a consensus sequence associated with calcium binding: (D/N)X(D/N)(E/Q)X_m(D/N)*X_n(Y/F) (where m and n are variable and an asterisk indicates -hydroxylation) (16). This subset of the EGF domain with a calcium-binding signature has been classified as cbEGF domain and is well conserved among a variety of proteins (Fig. 1B). Tandemly repeated cbEGF domains form a rodlike helix structure, contributing to the structural integrity of the proteins. Not only important as structural elements, cbEGF domains are also known to mediate protein-protein interactions (17).

The deduced amino acid sequence of the DANCE transcript also contains two putative Asn glycosylation sites and an RGD motif in the first cbEGF domain, which is known to be a common ligand sequence motif of integrin ligands (Fig. 1A) (18).

Cloning of Human and Rat DANCE-- The rat homologue of DANCE was obtained by screening a rat heart cDNA library and by 5'-RACE. Partial sequence of a human DANCE homologue was found in EST data bases. By complete sequencing, this human clone turned out to contain the complete coding region. Alignment of mouse, rat, and human DANCE amino acid sequences reveals that DANCE is quite well conserved among species (Fig. 1C). Amino acid identity between mouse and rat and between mouse and human is 98 and 94%, respectively. The RGD motif and putative Asn glycosylation sites are conserved among species.

DANCE, S1-5/T16, and UPH1/H411 Comprise a New EGF-like Protein Family-- A sequence homology search against the protein and nucleotide data base revealed that two proteins in particular are highly homologous to DANCE (Fig. 1D). S1-5 (human) was cloned from fibroblasts of a patient with Werner's syndrome and was reported to be overexpressed in Werner's syndrome and to stimulate DNA synthesis (19). T16 is a rat homologue of S1-5 and has 93% amino acid identity with S1-5. UPH1 (human) and H411 (Chinese hamster) are unpublished sequences found in the GenBank^TM data base. UPH1 and H411 share 93% amino acid identity. All of these have six cbEGF domains, one at the N-terminal and five tandemly repeated in the middle of these molecules. These proteins have significant homology both in the cbEGF repeat (human DANCE versus UPH1, 54%; human DANCE versus S1-5, 48%; S1-5 versus UPH1, 57%) and C-terminal domain (human DANCE versus UPH1, 53%; human DANCE versus S1-5, 50%; S1-5 versus UPH1, 53%). Therefore, these proteins are considered to comprise a new EGF-like protein family, which we propose to call the "DUS" (DANCE, UPH1, and S1-5) family. Among the DUS family, only DANCE has an RGD motif in the first cbEGF domain.

C-terminal domains of DANCE and other DUS family members have a weak homology with those of fibulin-1C, fibulin-1D, and fibulin-2, which are extracellular matrix proteins with cbEGF domain repeats (20-22) but much larger than DUS family members (Fig. 1E). The C-terminal domain of fibulin-1C is reported to be involved in binding to nidogen (23), but function for those of fibulin-1D and fibulin-2 are not known. The multiple domain structure of DANCE protein suggests that it may bind to multiple receptors or ligands, e.g. integrin and extracellular matrix, or integrin and other cell surface receptors.

Recombinant Expression of DANCE Protein-- COS7 cells and 293T cells were transiently transfected with a full-length rat DANCE cDNA in an expression plasmid vector. Cell lysate and conditioned media of transfected cells were analyzed by Western blot using a polyclonal antibody raised against C-terminal polypeptide of rat DANCE (Fig. 2). DANCE protein was detected in both cell lysate and conditioned media, demonstrating that DANCE is a secreted protein. The size of the expressed DANCE protein is approximately 66 kDa. No dimer or trimer of DANCE was observed even with nonreducing gel electrophoresis and Western blot (data not shown).

View larger version (22K): [in this window] [in a new window]   Fig. 2.   Recombinant protein expression of rat DANCE and detection by immunoblot. DANCE expression plasmid or vector plasmid were transfected into COS7 or 293T cells. After transfection, cells were cultured in serum-free media for 24 h. The media and cell lysate were analyzed by Western blot using anti-DANCE polyclonal antibody. Several artifactual bands that cross-reacted with anti-DANCE antibody are seen in 293T cell lysates.

Expression Profile of DANCE in Human Tissues and Chromosomal Mapping-- Northern blot analysis using poly(A) RNA from various human tissues and a full-length human DANCE cDNA probe revealed that the major transcript of 2.6 kilobase pairs is expressed mainly in heart, ovary, and colon (Fig. 3). Less expression is found in most of the tissues examined. However, it is undetectable in brain, liver, thymus, prostate, and peripheral blood leukocyte. This expression profile is largely different from that of S1-5, in which expression in heart is weak (24). Since each DUS family member is extremely conserved among species, it is likely that each member of the DUS family may have a distinct developmental or physiological role, despite the homology between the members.

View larger version (67K): [in this window] [in a new window]   Fig. 3.   Northern blot analysis of human DANCE mRNA. The blots from various adult human tissues containing 2 µg of poly(A) RNA in each lane were probed with human DANCE cDNA. The bottom panel shows the same blot hybridized with -actin. PBL, peripheral blood leukocytes.

The chromosomal localization of the human DANCE gene was determined by fluorescence in situ hybridization using full-length human DANCE cDNA as a probe. All 24 metaphase cells examined showed a specific hybridization signal with twin spots at 14q32.1 (Fig. 4, A and B). No known genetic disorder has thus far been mapped to this locus. The S1-5 locus is chromosome 2p16 (24), and fibulin-1 and fibulin-2 genes have been mapped on chromosomes 22q13.2-13.3 and 3p24-25, respectively (25, 26); thus, DUS family members and distantly related family genes have independent loci.

View larger version (96K): [in this window] [in a new window]   Fig. 4.   Chromosomal mapping of the human DANCE gene. Specific hybridization signals indicated by arrows are shown in A, which were identified as chromosome 14q32.1 with Q-banding of the same chromosomes in B.

DANCE Transcript Localization in Mouse Embryo, Adult Arteries, and Atherosclerotic Lesions-- DANCE expression in mouse development was studied by in situ hybridization. At 8.5 dpc, expression was restricted to endothelial cells of maternal placenta (Fig. 5, A and A'), whereas almost no signal was detected in embryonic tissues (data not shown). In 9.5-dpc embryos, migrating neural crest cells and the pericardium express DANCE (Fig. 5, B and B'). Branchial arch mesenchymal cells derived from neural crest cells continue to express the DANCE gene. No expression is detected in the neural tube. At 12.5-dpc, the cardiac outflow tract and aorta display extremely strong expression, whereas less but significant expression is detected in several other tissues (Fig. 6, A and A'). In the heart, the DANCE transcript is present in endocardial cushion tissues, which derive from endocardial cells by an epithelial-mesenchymal transition (27) (Fig. 5, C and C'). Together with DANCE expression in neural crest cells, these observations suggest that one developmental function of DANCE may be to play a role in epithelial-mesenchymal transitions. Head mesenchyme, intersomitic tissues, and several other mesenchymal tissues also express DANCE (Fig. 6). In the aorta, DANCE expression is seen both in endothelial cells and in smooth muscle cells. In 14.5-dpc embryos, some neural crest-derived tissues such as head mesenchyme, cardiac outflow tract, and sympathetic ganglia continue to express DANCE, but some neural crest tissues such as adrenal gland do not (data not shown). Mesenchymal tissues in proximity to developing cartilages also express DANCE mRNA (data not shown). The aorta still exhibits the strongest expression of DANCE at 14.5 dpc.

View larger version (110K): [in this window] [in a new window]   Fig. 5.   Expression of DANCE mRNA in several representative mouse embryonic sections as shown by in situ hybridization, viewed in dark field (A, B, and C) and bright field (A', B', C'). A and A', high magnification view of maternal placenta at stage 8.5 dpc. DANCE expression is found in endothelial cells. B and B', transverse section of rostral region of 9.5-dpc embryo. Neural crest cells migrating to branchial arches are indicated by arrowheads. Pericardium is indicated by an open arrowhead. C and C', heart region of 12.5-dpc embryo. DANCE expression in outflow tract (arrow) and endocardial cushion tissue (arrowhead) are shown. Magnification: × 200 (A and A'), × 40 (B and B'), × 40 (C and C').

View larger version (79K): [in this window] [in a new window]   Fig. 6.   Expression of DANCE mRNA in a 12.5-dpc embryo as shown by in situ hybridization with dark field (A) and bright field (A') views. a, aorta; d, duodenum; h, heart; li, liver; m, mandible; mv, mesenteric vessels; nt, neural tube; o, cardiac outflow tract; tv, third ventricle; v, vertebra. Magnification, × 40.

In adult aorta, DANCE expression is largely diminished. However, intense focal expression is found at intercostal branching points in the thoracic aorta (Fig. 7, A and A'). This observation is interesting in relevance to atherosclerosis, because alternation of hemodynamic stress at branching regions has been pointed out to induce atherogenesis at these regions (28, 29).

View larger version (90K): [in this window] [in a new window]   Fig. 7.   Expression of DANCE mRNA in normal adult mouse aorta and in atherosclerosis lesions of LDL receptor-deficient mice. Longitudinal section of normal adult thoracic aorta (A and A') shows robust expression only in the intercostal branching points (arrowheads). Transverse section of the thoracic aorta of LDL receptor-deficient mice (B and B') shows re-expression of DANCE mRNA in the endothelial cells overlaying fibrous plaques of atherosclerotic lesion (arrowheads), whereas only little expression is seen in the normal region (open arrowhead). Original magnification, × 40 (A and A'), × 40 (B and B').

Accordingly, DANCE expression was studied in atherosclerotic vessels using LDL receptor-deficient mice fed with a high cholesterol diet (30). Endothelial cells overlying the plaques exhibited a significant increase in DANCE mRNA expression (Fig. 7, B and B', arrowhead) compared with normal regions of the same vessel (open arrowhead).

Augmented Expression of DANCE in Balloon-injured Vessels-- To examine if DANCE is re-expressed in other settings of pathological vascular remodeling, DANCE transcript expression in balloon-injured rat carotid arteries and aortae was studied (31). Northern blot analysis revealed that DANCE mRNA expression was markedly increased following balloon injury with the highest levels seen at 14 days, coinciding with decreasing smooth muscle cell replication (Fig. 8). By in situ hybridization carried out with en face preparations of vessel segments, augmented expression of DANCE mRNA is observed in both endothelial cells and smooth muscle cells (Fig. 9). Endothelium from normal aorta shows no detectable expression (Fig. 9A). 8 days following endothelial wounding (Fig. 9B), little expression is seen at the leading edge of the regenerating endothelium (dotted line), but increased expression is evident further behind the leading edge where cells return to quiescence. Proliferating smooth muscle cells that migrated from the underlying media onto the luminal surface of the denuded aorta at 8 days after injury express DANCE mRNA (Fig. 9D), but even higher levels of expression are seen at 14 days when smooth muscle replication is decreasing (Fig. 9E). This observation suggests that DANCE may affect cell growth as a "brake" in autocrine or paracrine manner when proliferation should stop.

View larger version (70K): [in this window] [in a new window]   Fig. 8.   Expression of DANCE mRNA in the rat carotid artery following balloon catheter injury. The left carotid artery was denuded with a 2F balloon catheter as described (31). Total RNA was isolated from carotid arteries 6 h and 3, 7, and 14 days after balloon injury and from normal vessels with the endothelium removed (nor. carotid EC) (33). RNA was analyzed by Northern blotting using DANCE cDNA as probe. An RNA loading control with ethidium bromide-stained 28 S rRNA is shown below.

View larger version (176K): [in this window] [in a new window]   Fig. 9.   Expression of DANCE mRNA in endothelial cells and smooth muscle cells of injured rat aortae. In situ hybridization was carried out on en face preparations of vessel segments as recently described (13) using DANCE sense (C and F) and antisense probes (A, B, D, and E). A, normal endothelial cells. B and C, endothelial cells 8 days after denudation. The leading edge of regenerating endothelium is indicated by the dotted line in B. Proliferating smooth muscle cells that migrated to intima at 8 days after injury (D) and 14 days after injury (E and F). Magnification, × 400.

DANCE Protein Promotes Endothelial Cell Attachment through RGD-Integrin Interaction-- Because DANCE has an RGD motif, which is known as an integrin-binding motif, we studied whether it has a role in cell attachment. For this purpose, recombinant human DANCE protein was purified from the conditioned medium of transfected 293T cells. HUVECs spread on DANCE coated microtiter plates (Fig. 10A), whereas no cells spread on BSA-coated plates (Fig. 10B). This attachment to DANCE protein was dose-dependent (Fig. 10C). A synthetic peptide GRGDSP completely inhibited spreading and attachment of HUVECs to DANCE, whereas the control peptide GRGESP showed no inhibition (Fig. 10D). This RGD dependence of the cell attachment activity of DANCE suggests that the attachment is mediated by binding of cell surface integrins to DANCE protein.

View larger version (88K): [in this window] [in a new window]   Fig. 10.   DANCE mediates adhesion of endothelial cells through binding to integrins. HUVECs were allowed to attach on DANCE-coated wells (A) or BSA-coated wells (B). Each well was coated with 8 µg/ml of respective protein solution. C, HUVECs were incubated on 96-well plates that were precoated with various concentrations of DANCE protein or BSA protein solution. After washing, the relative number of attached cells were determined using wells without washing as standards. Data were obtained as quadruplicate, and mean ± S.D. values are shown. D, HUVECs preincubated with various concentrations of GRGDSP or GRGESP peptides were allowed to attach to wells coated with 8 µg/ml DANCE. Wells were washed as above, and the relative numbers of attached cells were determined.

Considering the multiple domain structure and robust expression in embryonic arteries, atherosclerotic lesions, and balloon-injured arteries, an intriguing possibility is that DANCE may contribute to vascular remodeling via interaction with integrins and other extracellular molecules.

	ACKNOWLEDGEMENTS

We thank Dr. S. Steinberg and Dr. W. Palinski for kindly providing us low density lipoprotein receptor-deficient mice fed with a high cholesterol diet. We thank Dr. T. Nakano (Department of Molecular Cell Biology, Research Institute for Microbial Diseases, Osaka University), Dr. H. Tada, Dr. S. Shibayama (Ono pharmaceutical, Japan), and Dr. J. McCoy (Genetics Institute) for helpful advice in constructing and screening of the signal sequence trap library and Dr. S. Evans and Dr. J. Chen (Department of Medicine, University of California at San Diego) for many suggestions in writing the manuscript. We also thank N. Tomikawa and S. Nomura for excellent technical assistance.

	FOOTNOTES

^* The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

The nucleotide sequence(s) reported in this paper has been submitted to the GenBank^TM/EMBL Data Bank with accession number(s) AF112151 (mouse DANCE), AF112152 (rat DANCE), and AF112153 (human DANCE).

^d Supported by Japan Society for the Promotion of Science Research Fellowships for Young Scientists and by an American Heart Association Western Affiliate Postdoctoral Fellowship.

^e A National Institutes of Health Individual National Research Service Award of National Institutes of Health recipient.

^j To whom correspondence should be addressed: Dept. of Medical Chemistry, Faculty of Medicine, Kyoto University, Yoshida Konoe-cho, Sakyo-ku, Kyoto 606, Japan. Tel.: 81-75-753-4371; Fax: 81-75-753-4388; E-mail: honjo@mfour.med.kyoto-u.ac.jp.

² This program is available on the World Wide Web.

	ABBREVIATIONS

The abbreviations used are: EGF, epidermal growth factor; cbEGF, calcium-binding EGF-like; RACE, rapid amplification of cDNA ends; BSA, bovine serum albumin; HUVEC, human umbilical vein endothelial cell; dpc, day(s) postcoitum.

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS AND DISCUSSION REFERENCES